Historical note and terminology

Currently, malformation of cortical development can be categorized as follows: (A) cortical dysplasia, which encompasses the full spectrum of neuronal migration disorders or malformations of cortical development, ranging from the subtlest to the most severe; (B) CNS structural lesions associated with tuberous sclerosis complex; (C) Sturge-Weber syndrome, or encephalotrigeminal angiomatosis; (D) neurofibromatosis type II, which may be associated with meningio-angiomatosis; and (E) vascular malformations, especially cavernous hemangiomas and arteriovenous malformations as described by Houser and Vinters (24). Historically, malformation of cortical development in patients with intractable epilepsy was initially described by David Taylor and colleagues in 1971, when the term "focal cortical dysplasia" was established. They noted dysplastic neurons and aberrant and disoriented processes associated with a loss of normal cortical lamination and large, poorly differentiated cells that displayed features of both neurons and glia (balloon cells) (51). The pathological similarity of cortical dysplasia to the cortical hamartomata associated with tuberous sclerosis complex was already mentioned. The classification of focal cortical dysplasia has evolved over time. Differentiation of focal cortical dysplasia type I and type II was made with further detailed pathological description in 2004 (43). The most recent amendment was done in 2011 (05), with an introduction of focal cortical dysplasia type III, which is associated with other brain malformations and destructive lesions. Type IIIA occurs in combination with hippocampal sclerosis, type IIIB is seen with epilepsy-associated tumors, and type IIIC is found adjacent to vascular malformations, whereas focal cortical dysplasia type IIID can be diagnosed in association with epileptogenic lesions acquired in early life (ie, including hypoxic-ischemic injury or encephalitis) (05; 53). As an example, Sturge-Weber syndrome causes intractable epilepsy in children, often due to cortical malformation in association with leptomeningeal angiomatosis (which would be classified as focal cortical dysplasia type IIIC) (37; 40).

Of note, the current classification is based purely on pathological characteristics, and it does not utilize any molecular or genetic information. As an example of pathological-genetic correlation, balloon cells that are typically seen in focal cortical dysplasia type IIB can be observed in cortical tubers from patients with tuberous sclerosis complex and hemimegalencephaly, which suggests a common pathway in pathophysiology, such as the crucial involvement of the mTOR pathway during cellular proliferation and differentiation (03). On the contrary, focal cortical dysplasia type I seemed a heterogeneous group of disorders resulting from late, postmigrational insults to the developing cerebral cortex (22). Of note, glioneuronal tumors, including gangliogliomas and dysembryoplastic neuroepithelial tumors, have dysplastic neurons like those seen in cortical dysplasia in addition to glial tumors. They also demonstrate evidence for over activation of the mTOR pathway (07). Efforts have been made to revise the focal cortical dysplasia classification to integrate clinic-pathological, histopathological, and genetic classification system, and such revision will likely be implemented in the near future (41; 04).

Clinical manifestations

Focal seizures are the most common manifestation of lesions due to malformation of cortical development. Based on an international multicenter study including more than 500 children who underwent epilepsy surgery, 42% of them had cortical dysplasia (of those, 10% had hemimegalencephaly) (23). Some may appear as generalized seizures, as for epileptic spasms in children with focal cortical dysplasia (47), yet investigation for focal lesion is crucial in determining early surgical intervention. An accurate incidence of epilepsy in focal cortical dysplasia is difficult to determine because mild or asymptomatic focal cortical dysplasia may go undiagnosed. However, in terms of epilepsy in focal cortical dysplasia, seizures occur in early life with shorter time to epilepsy surgery compared to temporal lobe epilepsy with hippocampal sclerosis or other etiologies. In fact, focal cortical dysplasia is the most frequent histopathology in children and the second or third most common etiology in adults. In one study, including children who had undergone respective surgery for drug-resistant epilepsy, about 15% presented with status epilepticus, and 30% of patients had a history of infantile spasms (34). Medication-resistant epilepsy is especially common in patients with the tuberous sclerosis complex. About 85% develop seizures, and two thirds have seizure onset in the first year of life. Close to 40% have a history of infantile spasms, and two thirds have pharmacoresistant epilepsy (09). Even higher prevalence of epilepsy (more than 90%) and infantile spasms (more than 50%) are seen in hemimegalencephaly (17). The above-mentioned lesions can be treated with a high degree of success with surgery; thus, it is important to consider this option when patients with these conditions become medically intractable.

Biological basis

Embryologically, focal cortical dysplasia emerges from partial failure of the later phases of neocortical formation. A failure of late cortical maturation could explain the presence of abnormally thickened gyri with indistinct cortical gray-white matter junctions. Developmental alterations during the late second or early third trimester would account for severe cortical dysplasia, such as hemimegalencephaly, whereas events occurring closer to birth might explain milder forms of focal cortical dysplasia (24).

Genetics and molecular biology of malformations of cortical development been studied extensively in the past 2 decades. mTOR signaling alteration has been described in tuberous sclerosis complex (39), focal cortical dysplasia II, glioneuronal tumors, and hemimegalencephaly (48; 13; 29; 14), suggesting common pathogenesis among these conditions. It is also important to note mTOR signaling is over-activated in hippocampal sclerosis and Rasmussen encephalitis, suggesting a significant role of the mTOR pathway in epileptogenesis (26). DEPDC5, a member of the GATOR complex, negatively regulates mTOR pathway at the level of mTORC1, and mutations of the gene may present as autosomal dominant familial focal epilepsy in association with focal cortical dysplasia IIA and focal cortical dysplasia IIB, including bottom-of-the sulcus dysplasia (48; 26). The study showed second-hit somatic mosaic mutation over a germline mutation in DEPDC5 can cause epilepsy associated with focal cortical dysplasia (44). BRAF, a serine-threonine kinase, is a major regulator of the MEK/MAPK cascade, another regulator of cellular proliferation and differentiation. Mutations in the BRAF gene have been associated with tumors, including glioneuronal tumors (03), and may contribute to similarity in clinical presentation and pathological findings among focal cortical dysplasia and glioneuronal tumors. Immune response and inflammation may also play roles in epileptogenesis in these conditions. Microglial activation is seen in focal cortical dysplasia (06), and inflammatory markers including ICAM-1, TNF-alpha, and NF-kappaB are seen in tubers in tuberous sclerosis complex (35). Another study reported the expression of complement factor C3 was significantly upregulated in TSC and FCD2B (21).

Although GABA causes hyperpolarization and inhibition of neurons in adulthood, a relatively elevated intracellular chloride concentration in immature neurons leads to depolarization and excitation (55). Abnormally increased expression of NKCC1(Na-K-Cl co-transporter, seen in immature neurons, concentrates chloride ions inside the cell) and abnormally low distribution of KCC2 (K-Cl co-transporter, the mature form extrudes chloride from the cell) are seen in focal cortical dysplasia, hemimegalencephaly, and gangliogliomas, which resembles expression pattern in immature cortex and may contribute to hyperexcitability and seizure development in these conditions (02). Also, other dysregulation of inhibitory and excitatory synaptic transmission, including reduction in GABAergic interneurons and increase in excitatory neurotransmitter receptors or channels (eg, glutamate receptors, calcium channels), are reported (55; 03). A study showed spontaneous pacemaker-like GABA receptor-mediated synaptic activity was reported in surgical specimens from patients with focal cortical dysplasia and hemimegalencephaly, suggesting the presence of abnormal cells facilitating pyramidal neuron synchrony leading to epileptogenesis (08). This is reminiscent of the findings by Wu and colleagues in hypothalamic hamartomas where pacemaker-like activity was seen in dysmature GABAergic neurons, with persistent NKCC1 expression and network excitation produced by both glutamate and GABA (56). Further understanding of these molecular and signal cascades will likely facilitate development of disease-modifying pharmacotherapy.

Differential diagnosis

Brain tumors can mimic focal cortical dysplasia in terms of clinical presentation. Seizures are the primary clinical manifestation of glioneuronal tumors, with seizure prevalence being close to 100% in dysembryoplastic neuroepithelial tumors and more than 80% in gangliogliomas (52). Focal lesions can be seen using neuroimaging studies in both focal cortical dysplasia and brain tumors, and their clinical presentation and neuroimaging findings may be difficult to differentiate. In fact, pathology frequently shows both focal cortical dysplasia and tumor cells after resection (50).

Diagnostic workup

Neuroimaging study is the mainstay of diagnostic workup. Tuberous sclerosis complex and hemimegalencephaly can be easily identified by conventional MRI. Tubers in tuberous sclerosis complex can be detected as early as 20 weeks of gestational age. Regarding focal cortical dysplasia, 80% or more of focal cortical dysplasia type II have positive structural MRI, as opposed to 30% to 70% of those with focal cortical dysplasia type I (33; 46). EEGs are mostly positive in showing epileptiform discharges in patients with focal cortical dysplasia, but the discharges are often diffusely distributed and not localized or lateralized to epileptogenic zones (19). Focal cortical dysplasia type I and type IIA have high prevalence (one third to one half) of nonconcordant EEG and MRI findings. PET hypometabolism can also be negative, especially in focal cortical dysplasia type I. Thus, a multimodal approach (such as FDG-PET/MRI co-registration) helps increase the sensitivity to detect the lesion, improving surgical outcome and reducing the need of invasive monitoring (33; 46).

Cortical dysplasia in infantile spasms is more detectable with PET than MRI (10) although the timing of imaging in relation to development can influence the MRI signal characteristics (47). Other advanced MRI techniques, such as diffusion tensor imaging, can be helpful in detecting epileptogenic lesions, as shown in tuberous sclerosis complex (59). A fully-automated morphometric analysis of MRI using machine learning has a potential complement human effort in the detection of focal cortical dysplasia (12). High-field MRI, such as 7T MRI, may be utilized in the future to increase the sensitively of the detection of such lesions (60; 54). Magnetoencephalography is a noninvasive method for detection of focal epileptic activity with high temporal and spatial resolution and can be coregistered with MRI. Magnetoencephalography clusters may localize either within lesions or in their surroundings or may localize to one surrounding area. Magnetoencephalography can guide the detection of occult focal cortical dysplasia and placement of invasive electrodes (22).

Management

Medical treatment. Epilepsy due to malformation of cortical development often progresses to pharmacoresistance. Antiseizure drugs, including carbamazepine, oxcarbazepine, lacosamide, lamotrigine, levetiracetam, topiramate, sodium valproate, and clobazam for focal seizures, have been used for conventional medical therapy. The development of disease-modifying therapy has been investigated as understanding of the pathophysiology of these conditions has improved. In a double-blind randomized clinical trial involving patients with tuberous sclerosis complex who remained refractory to multiple medications, everolimus, an mTOR inhibitor, produced significant seizure reduction compared to placebo (18). Long-term follow-up data will be obtained from the same cohort. A pilot clinical trial investigating the effect of everolimus in patients with tuberous sclerosis complex and focal cortical dysplasia was completed and pending report as of 2021 (www.clinicaltrials.gov). The role of mTOR inhibition in modifying epileptogenesis will likely be further investigated in focal cortical dysplasia, hemimegalencephaly, and other mTOR-related disorders, which will create stepping stones for precision medicine in the treatment of epilepsy. Although mTOR inhibitors cannot reverse the already established cortical malformation, they may impact the plasticity involved in the epileptogenic process via modification of inflammatory signaling, interference with long-term potentiation, etc.

Surgical treatment. Focal cortical dysplasia is the most common pathological substrate for a child with drug-resistant epilepsy requiring surgery. It is one of the more challenging conditions to treat surgically, and its success rates are lower when compared to epilepsy surgery for benign brain tumors and temporal lobe epilepsy with hippocampal sclerosis (36). The relative poor prognosis is due to focal cortical dysplasia being often invisible on MRI, and even when visible, focal cortical dysplasia can exist well beyond the margins of the detectable radiographic lesion. The completeness of surgical resection is invariably associated with the extent of seizure-freedom achieved from this operation.

Following a multidisciplinary meeting at which the child’s history and physical examination, seizure semiology, EEG data, imaging data, and neuropsychological evaluation are carefully reviewed, a consensus decision would be made as to whether the child is a candidate for epilepsy surgery or not. Careful consideration of the patient’s and family’s values and preferences are required in order to make personalized surgical treatment recommendations (16). This is especially important if the epileptogenic region overlaps with key functional areas that may result in an unacceptable neurologic deficit if a complete resection of the lesion is carried out.

Intraoperative navigational systems have gained widespread utility in epilepsy surgery for lesion localization and craniotomy planning. This is especially important in cortical dysplasia given the pathology is not visually detectable in its entirety during surgery. In addition, it allows for visual pattern recognition of the underlying gyral and sulcal anatomy using 3D image renditions of the cortical gyri. However, this does not substitute the need for anatomical knowledge and functional neuroanatomy. Image fusion allows for the integration of preoperative anatomical MRI data with EEG source localization, functional MRI, MEG, PET, ictal SPECT, or diffusion tensor imaging. This allows the surgeon to define the 3-D resection boundaries following supplemental data from intra- or extra-operative electrocorticography data. EEG recordings from the brain have greater spatial resolution than EEG recordings from the scalp; however, brain EEG will always be much more spatially limited, and coverage of the brain will be incomplete.

When preoperative data are insufficiently discordant, MRI is negative, or the proposed resection involves or is adjacent to eloquent brain cortex, an invasive evaluation may be indicated for brain mapping or delineation of the seizure onset zone and irritative zone. Subdural electrodes, in the form of grids and strips, are best for recording from the lateral surface of the brain when brain mapping is essential, in particular for language mapping. Depth electrodes are particularly useful for recording limbic structures, including the hippocampus, cingulate gyrus, orbitofrontal, and insular regions. In our experience at UCLA, the decision regarding the resection boundaries generally is weighted more towards the anatomical abnormality rather than ECoG findings. Commonly, we have a hybrid approach to invasive evaluation as we combine the use of depth electrodes with surface grid and strip electrodes.

Careful consideration must be given to the amount of hardware placed in the intracranial space. Although the morbidity of an invasive evaluation is low at experienced pediatric epilepsy centers, the risks include CNS infections, intracranial hemorrhages (eg, epidural, subdural and intraparenchymal hemorrhages), malignant cerebral edema, mass effect, and cerebrospinal fluid leak (15; 58; 38). Electrode fractures have also been reported. Mortality is exceedingly rare at specialized centers (49).

Whether or not the child undergoes an invasive EEG evaluation, the completeness of surgical resection, as defined by the MRI lesion and the EEG data, of the affected area will depend on a risk versus benefit assessment, especially when the resection margins are adjacent to or involve eloquent cortex. The risk of residual seizures from an incomplete resection versus the risk of neurologic deficits resulting from surgery must be carefully balanced. Tactile differentiation requires experience, but it is an important tool that can be used by an epilepsy surgeon to differentiate normal tissue from focal cortical dysplasia. Visual differentiation is extremely challenging, and focal cortical dysplasia is typically visually indistinguishable from normal cortex. Although the resolution of an intraoperative MRI may be lower compared to preoperative imaging, there may be a role for it to ensure full resection of the anatomical lesion (32).

Surgical strategies in focal cortical dysplasia can include lesionectomies, lobectomies, and multi-lobar or hemispheric resections. Patients with focal cortical dysplasia type II are less likely to have a normal MRI as opposed to patients with focal cortical dysplasia type I. Younger patients and those with focal cortical dysplasia type I generally undergo lobar or multi-lobar resections as the pathology is commonly widespread (31). If the child is under the age of 2 years and has a corresponding contralateral loss of fine-motor strength and a homonymous hemianopsia, a hemispherectomy may be a considered if the epileptogenic zone is widespread and lateralized to one hemisphere. This is in contrast to older children or those with focal cortical dysplasia type II who are less likely to undergo a multi-lobar resection. Here, the boundaries of the resection tend to be limited by eloquent cortex. In fact, resections limited by eloquent cortex lead to one of the most frequent causes of unsuccessful surgical treatment (11). For patients with a positive transmantle sign, our practice at UCLA is to fully excise the “tail” of the malformation completely to the level of the ventricle. Postoperative mortality is low (49). The most common transient postoperative complications include infections and mild neurologic deficits. Permanent neurologic deficits or the development of hydrocephalus is uncommon (30).

The task of careful patient selection at the multidisciplinary conference cannot be overstated. The concordance of presurgical data and the ability to fully resect the presumed epileptogenic cortex is likely the greatest predictor of outcome. Seizure outcomes following resective epilepsy surgery are highly variable, and the predictive value of pre- and postoperative factors associated with seizure freedom are inconsistent (45). A meta-analysis based on primarily single-center cohort studies demonstrated that partial seizures, temporal resections, positive MRI demonstrating a focal cortical dysplasia, focal cortical dysplasia type II (as opposed to focal cortical dysplasia type I), and completeness of resection with respect to anatomical or electrographic abnormalities all predicted a greater likelihood of postoperative seizure freedom (31; 45).

In focal cortical dysplasia type II, scalp EEG and intraoperative and extraoperative electrocorticography (ECoG) or stereo EEG (SEEG) may show a characteristic pattern as rhythmic epileptiform discharges that spatially correlate with the anatomic extent of the lesion (22). Studies have shown high-frequency oscillations may be better marker of epileptogenic zones than conventional neurophysiological findings, including spikes and slowing (20). In retrospective studies, complete removal of brain areas showing high-frequency oscillations was associated with good seizure outcome (27; 57). Focal cortical dysplasia may have widely distributed epileptogenic zones, sometimes with negative neuroimaging studies. High-frequency oscillations can be helpful for determining the resection margins during intraoperative monitoring (57; 25) and extra-operative monitoring (01; 42) to improve seizure outcome in focal cortical dysplasia. However, a prospective multicenter study using high-frequency oscillations failed to demonstrate the removal of high-frequency oscillation-generating regions was associated with good postoperative seizure-free outcome, suggesting further standardization of the analytic methodology and validation in clinical trials are needed to utilize high-frequency oscillations in clinical practice (28).