Genotype‐phenotype correlations in focal malformations of cortical development: a pathway to integrated pathological diagnosis in epilepsy surgery

Abstract

Malformations of cortical development (MCD) comprise a broad spectrum of developmental brain abnormalities. Patients presenting with MCDs often suffer from drug‐resistant focal epilepsy, and some become candidates for epilepsy surgery. Their likelihood of achieving freedom from seizures, however, remains uncertain, and depends in a major part on the underlying pathology. Tissue samples obtained in epilepsy surgery form the basis of definite histopathological diagnosis; however, new molecular genetic methods have not yet been implemented in diagnostic processes for MCD cases. Furthermore, it has not been completely understood how the underlying pathology affects patients’ outcomes after epilepsy surgery. We performed a systematic literature review of studies describing both histopathological and molecular genetic findings in MCD, along with studies on epilepsy surgery outcomes. We aimed to correlate the genetic causes with the underlying morphological abnormalities in focal cortical malformations and to stress the importance of the underlying biology for patient management and counseling. From the summarized findings of multiple authors, it is obvious that MCD may have a diverse genetic background despite a similar or even identical histopathological picture. Even though most of their molecular genetic findings converge on various levels of the PI3K/AKT/mTOR pathway, the exact mechanisms underlying MCD formation have not yet been completely described or indeed how this pathway generates a diverse range of histological abnormalities. Based on our findings, we therefore propose that all patients diagnosed and operated for drug‐resistant epilepsy should have an integrated molecular and pathological diagnosis similar to the current practice in brain tumor diagnostic processes that might lead to more accurate diagnosis and effective stratification of patients undergoing epilepsy surgery.

Introduction

Malformations of cortical development (MCD) comprise multiple disease entities associated with abnormalities in the structure of the cerebral cortex and subcortical structures, a significant proportion of which leads to the development of drug‐resistant epilepsy, often manifesting in the first months to years of life. In carefully selected cases, epilepsy surgery may lead to significant improvements in seizure frequency, or even complete seizure freedom, and MCD represent the most prevalent cause of drug‐resistant epilepsy in pediatric epilepsy surgery cases 19. In the recent extensive reviews published by Englot and colleagues 13, 14 56% of the pediatric patients undergoing epilepsy surgery for extra‐temporal lobe epilepsy have reached seizure freedom (Engel class I outcome), as compared to 76% of children with temporal lobe epilepsy 14. Many other authors have established outcomes in pediatric epilepsy surgery cohorts and aimed to identify predictive factors associated with seizure‐free status after the surgery in FCD and TSC cases, which include completeness of resection 15, 28, 29, FCD type II histopathology 31, more restricted EEG abnormities 28, younger age at surgery and unilobar distribution 15. In other large pediatric epilepsy surgery cohorts, however, the authors did not observe a significant relationship between MCD type and outcome of surgery 51. Even more strikingly, the outcomes for MCD patients tend to not change remarkably over many years in established epilepsy surgery programs, and a significant proportion of patients do not reach freedom from seizures or from antiepileptic drugs 20, 33. These findings are supported by the most recent, and to our knowledge, most extensive study on patients undergoing epilepsy surgery that combines rigorous histopathological work‐up with detailed information on clinical status and outcome 8. The study has shown that roughly 40% of the patients from established epilepsy surgery centers failed to reach freedom from seizures one year after surgery. However, no molecular genetic studies targeting possible somatic variants in genes associated with MCD were performed in this study.

In conclusion, despite undoubted improvements in surgical techniques, patient selection and referral to specialized centers, there remains a significant proportion of patients who fail to achieve freedom from seizures. Furthermore, none of the studies have considered the underlying molecular genetic changes as a potential factor influencing patient selection as epilepsy surgery candidates and their chances to reach seizure freedom. This is in striking contrast to the amount of emerging data on genetic background in various MCD syndromes identified so far, mostly by means of next generation sequencing (NGS) techniques. Molecular genetic methods have yet not been recognized as a standard part of surgical neuropathology diagnostic process in MCD cases 7; again, this contrasts with its unquestionable role in brain tumor diagnostics where specific molecular genetic changes constitute a cornerstone of diagnosis 44 and provide guidance to estimate prognosis. Our knowledge of how the genetic background of MCD relates to surgical outcome of MCD patients is scattered in the literature, usually consisting of case reports or reports of small patient series that have focused on clinical and radiological findings. Moreover, there is little data to indicate how the underlying pathological changes relate to the genetic alterations. This is an important step as it integrates the genetic and cellular findings and can inform the diagnosis of patients undergoing epilepsy surgery. In the process of establishing a targeted next generation sequencing panel for MCD surgical neuropathological diagnostics in our centers, we looked for high‐quality systematic reviews and original reports on the role of molecular genetic changes in genes associated with MCD formation. In this review, we summarize the published findings on the topic, the histopathological and the molecular genetic findings in MCD, with an aim to provide clinicians and neuropathologists, particularly in epilepsy surgery centers, with guidance for their future molecular genetic studies, both in clinical and research settings.

Methods

We searched for high‐quality review articles in PubMed published between 2000 and 2017 with full‐text in English available from established journals in the field; for the search build‐up, we used MeSH terms “malformation of cortical development” and “review.” We undertook this approach in order to overcome the difficulties of identifying rare pathological descriptions in otherwise scattered clinical reports. Then, we selected reviews that were relevant to the topic (MCD associated with epilepsy, amenable to surgical treatment). From the published reports, we looked for original reports on genes involved and selected those with a detailed report on the histopathological findings. We included reports with a description of somatic pathological genetic variants and polymorphisms in brain tissue samples and also reports of focal MCD with detected germline variants. Papers without a clear description of histopathological findings were excluded. The molecular, genetic and histopathological findings, along with the methods used in the reports are summarized in Supplementary materials.

Results

The histopathological picture of focal malformations of cortical development

Focal cortical dysplasia represents the most common etiology of drug‐resistant focal epilepsy amenable to surgery in children and the third most common etiology in epilepsy surgery cases in adults 8, 19. The ILAE classification recognizes three major groups of FCD; type I with abnormalities of cortical architecture, type II with abnormalities of cytology and type III adjacent to other brain lesions 9. Type Ia FCD is characterized by an abnormal radial cortical lamination, while type Ib by an abnormal tangential cortical lamination, and type Ic with a mixture of abnormal radial and tangential lamination. Abnormal cells are the defining feature of FCD type II; dysmorphic neurons of FCD IIa and dysmorphic neurons and balloon cells of FCD IIb. The typical attributes of dysmorphic neurons are: enlarged neuronal nucleus and cell body, peripherally displaced and perinuclear Nissl substance, and cytoplasmic accumulation of phosphorylated and non‐phosphorylated neurofilaments. In contrast, balloon cells are easily distinguished by their large cell body and even eosinophilic cytoplasm with no neurofilaments present. Type III FCD is characterized by abnormal cortical lamination (similar to type I) but accompanies other brain lesions; type IIIa adjacent to hippocampal sclerosis in temporal lobe, type IIIb in the vicinity of glial or glioneural tumour, type IIIc accompanying vascular lesions and type IIId adjacent to an acquired lesion (e.g. inflammatory, ischemic or traumatic).

In recent years, the new sequencing methods have vastly expanded the possibilities of detecting low‐level somatic mutations in FCD brain tissue samples. This has led to discoveries of pathogenic variants in multiple genes that might contribute to the formation of focal cortical dysplasia. Multiple studies have reported somatic pathogenic variants in genes of PI3K/AKT/mTOR pathway in FCD type II tissue samples 23, 40, 41, 48, 50; however, reports on genetic background of FCD type I have been fairly scarce 5, 10.

Summary of MRI features of surgically amenable MCD

Magnetic resonance imaging (MRI) constitutes the main diagnostic modality for MCD. The spectrum of severity of MRI changes ranges from no observable features and subtle changes, especially prevalent in FCD type I 1, 26, 30, to obviously extensive involvement of the entire hemisphere. The distinguishing features of FCD type I include cortical thinning, abnormal signal on T2w, T1w and FLAIR sequences, lobar hypoplasia and blurring of grey and white matter 1, 26, 30. FCD type II is characterized by increased cortical thickness, blurring of the grey and white matter, T2w and T1w signal abnormalities and abnormal gyral pattern; this applies to both FCD type IIa and IIb, and FCD type IIb typically harbors also the transmantle sign (a T2w hyperintense connection between the subcortical white matter and the lateral ventricle) 1. FCD type III, accompanied by other structural brain pathology, is mostly recognized by the original pathology itself (e.g. brain tumor, hippocampal sclerosis) with FCD surrounding the lesion 26. It is important to stress that these features can only be observed in maturely myelinated brain; the MRI features of FCD differ before brain myelination is completed (approximately at the age of one year), and FCD lesions may appear as regions hypointense in T2w sequences that disappear with ongoing myelination 12. In hemimegalencephaly, termed also dysplastic megalencephaly, variable enlargement of the entire or a part of one hemisphere occurs, possibly with atypical pachygyria, accompanied by some features of FCD (e.g. blurring of grey and white matter, T2w signal abnormalities) 17.

MRI findings in TSC encompass a broad spectrum of dysplastic changes, in addition to the tumor and tumor‐like lesions of the subependymal giant cell astrocytoma (SEGA) and subependymal nodules (SEN). The typical MRI features of TSC include cortical and cortico‐subcortical tubers and white matter migration lines 53, in addition to features similar to or identical with isolated FCD 22.

Somatic variants in FCD type II

First attempts to detect genetic changes in FCD tissue samples were made in early 2000s, based on histopathological similarities between tubers of tuberous sclerosis complex and focal cortical dysplasia type IIb. Indeed, multiple authors have searched for common genetic background of TSC and FCD, with equivocal results; in some studies, they detected allelic variants in TSC1 and TSC2 genes 6, 46, other studies found no distinctive pathogenic variants in either TSC1 or TSC2 in their cohorts 18. The advent of next‐generation sequencing and new experimental methods to introduce suspected genetic variants in animal models (e.g. in utero electroporation, CRISPR/Cas9) has brought new insight into and provided solid evidence for the involvement of TSC1 and TSC2 in FCD 40. The group of Lim and colleagues performed deep hybrid capture and amplicon sequencing on brain tissue samples from 40 individuals and detected pathogenic variants in TSC1 and TSC2 in 12.5% of the patients (for both FCD type IIa and IIb cases). They have reported variants in TSC1 and TSC2 and also constructed a mouse model as a proof‐of‐principle of their role in FCD formation 41. Surprisingly, the first results of deep sequencing studies of tuber tissue samples did not show second‐hit mutations in either TSC1 or TSC2 to contribute to tuber formation in tuberous sclerosis lesions; a second‐hit mutation event was detected in a single sample among 46 tubers in a study by Qin and colleagues 55. However, later studies have brought contradictory findings; Martin et al 47 found second‐hit mutations in 35% of the tubers, as compared to the majority of renal angiomyolipomas and subependymal nodules (SEN) or subependymal giant cell astrocytomas (SEGA), another landmark lesion of TSC. Still, this study points to the fact that second‐hit mutations in tubers, as a specific type of MCD, are rather less frequent events than in typical tumors, and their role in tuber formation therefore remains enigmatic. This is further supported by the unexpected discovery of combined somatic and germline variants in TSC2 in two patients with hemimegalencephaly without any other features of TSC 11.

Given the evidence for the involvement of TSC1 and TSC2in FCD, it comes as no surprise that pathogenic variants in MTOR, which is downstream of TSC1 and TSC2, have been detected both in FCD IIa and FCD IIb 42, 48, 50, 52. Lim's group first detected somatic mutations in MTOR in FCD type IIa and IIb tissue samples in 15.6% of the patients using deep WES and hybrid capture and amplicon sequencing. Subsequently, they performed in utero electroporation of the detected mutation to prove its pathogenicity; this lead to the development of spontaneous seizures in mice, along with cytological abnormalities similar to those observed in human FCD 42. MTOR variants turned out to be especially prevalent in FCD samples, as described by multiple other authors (see Table 1, 2) 49, 50, 52.

Table 1.

Overview of somatic mutations detected in MCD. Abbreviations: FCD = focal cortical dysplasia; HMEG = hemimegalencephaly.

Neuropathological diagnosis	Gene	Mutation identified (somatic unless specified otherwise)	Paper	MCD type	MCD extent
FCD IIa	PIK3CA	p.His1047Arg	Jansen LA, Mirzaa GM, Ishak GE, et al Brain. 2015	FCD	focal
		p.Glu545Lys (c1633G>A)	Lee JH, Huynh M, Silhavy JL, et al Nat Genet. 2012	HMEG	hemispheric
	AKT3	p.Glu17Lys	Jansen LA, Mirzaa GM, Ishak GE, et al Brain. 2015	HMEG	hemispheric
		p.Glu17Lys (c.49C>T)	Lee JH, Huynh M, Silhavy JL, et al Nat Genet. 2012	HMEG	hemispheric
		p.Glu17Lys (c.49C>T) in 1, 1q trisomy in 2	Poduri A, Evrony GD, Cai X, et al Neuron 2012	HMEG	hemispheric
	MTOR	c.4487T>G; p.W1456G	Leventer RJ, Scerri T, Marsh AP, et al Neurology. 2015	hemispheric dysplasia	hemispheric
		c.7280T>C (p.Leu2427Pro); c.6577C>T (p.Arg2193Cys); c.1871G>A (p.Arg624His); c.5126G>A (p.Arg1709His)	Lim JS, Kim WI, Kang HC, et al Nat Med. 2015.	FCD	focal
		p.Leu1460Pro (c.4379T>C); p.Ser2215Phe (c.6644C>T); p.Ser2215Tyr (c.6644C>A)	Mirzaa GM, Campbell CD, Solovieff N, et al JAMA Neurology. 2016	FCD	focal
		c.6644C.A/p.Ser2215Tyr	Moller RS, Weckhuysen S, Chipaux M, et al Neurol Genet. 2016	FCD	focal
		p.Cys1483Tyr (c.4448C>T)	Lee JH, Huynh M, Silhavy JL, et al Nat Genet. 2012	HMEG	hemispheric
	DEPDC5	c.4187delC (p.Ala1396Valfs*78)	Mirzaa GM, Campbell CD, Solovieff N, et al JAMA Neurology. 2016	HMEG	hemispheric
		c.865C>T/ p.Gln289* (somatic) + c.856C>T/p.Arg286* (germline)	Ribierre, T., et al J Clin Invest. 2018	FCD	focal
	TSC2	c.4639G>A (p.Val1547Ile)	Lim JS, Gopalappa R, Kim SH, et al Am J Hum Genet. 2017	FCD	focal
	TSC1	c.610C>T (p.Arg204Cys)	Lim JS, Gopalappa R, Kim SH, et al Am J Hum Genet. 2017	FCD	focal
		c.64C>T (p.Arg22Trp)	Lim JS, Gopalappa R, Kim SH, et al Am J Hum Genet. 2017	FCD	focal
FCD IIb	PTEN	F278L (exon 8; c.834C>G)	Schick V, Majores M, Engels G, et al Acta Neuropathol. 2006	FCD	focal
	MTOR	c.6644C>T (p.Ser2215Phe); c.7280T>A (p.Leu2427Gln); c.5930C>A (p.Thr1977Lys); c.4348T>G (p.Tyr1450Asp); c.4447T>C (p.Cys1483Arg)	Lim JS, Kim WI, Kang HC, et al Nat Med. 2015.	FCD	focal
		p.Ser2215Tyr (c.6644C>A), p.Ala1459Asp (c.4376C>A), p.Leu1460Pro (c.4379T>C), p.Ser2215Phe (c.6644C>T)	Nakashima M, Saitsu H, Takei N, et al Ann Neurol. 2015; Lim JS, Kim WI, Kang HC, et al Nat Med. 2015.	FCD	focal
		c.4379T>C/p.Leu1460Pro; c.4375G>T/p.Ala1459Ser	Moller RS, Weckhuysen S, Chipaux M, et al Neurol Genet. 2016	FCD	focal
		c.6644C.A/p.Ser2215Tyr	Moller RS, Weckhuysen S, Chipaux M, et al Neurol Genet. 2016	FCD	focal
		c.6644C>T/p.Ser2215Phe	Moller RS, Weckhuysen S, Chipaux M, et al Neurol Genet. 2016	FCD	focal
	TSC1	c.64C>T (p.Arg22Trp)	Lim JS, Gopalappa R, Kim SH, et al Am J Hum Genet. 2017	FCD	focal
		453G>A, p.E78K, 549 G>A, p.A110T, 2415C>T p.H732Y (previously described in Jones)	Becker AJ, Urbach H, Scheffler B, et al Ann Neurol. 2002; Jones AC, Daniells CE, Snell RG, et al Hum Mol Genet 1997	FCD	focal
FCD I	AKT3	somatic trisomy of the 1q21.1‐q44 chromosomal region, encompassing the AKT3 gene	Conti V, Pantaleo M, Barba C, et al Clin Genet. 2015	FCD	focal
	DEPDC5	p.Arg422* (somatic) + p.Arg239* (germline)	Baulac S, Ishida S, Marsan E, et al Ann Neurol. 2015	FCD	focal

Table 2.

Overview of the germline mutations detected in MCD. Abbreviations: FCD = focal cortical dysplasia; TSC = tuberous sclerosis complex.

Neuropathological diagnosis	Gene	Mutation identified (germline unless specified otherwise)	Paper	MCD type	MCD extent
Tubers	TSC2	multiple germline	Qin W, Chan JA, Vinters HV, et al Brain Pathology. 2010	TSC	focal
		1864C>T R622W (2nd hit in addition to germline)	Qin W, Chan JA, Vinters HV, et al Brain Pathology. 2010	TSC	focal
	TSC1	737+1G>A, 827–828delCT, 1997+1G>A, 2347C>T	Qin W, Chan JA, Vinters HV, et al Brain Pathology. 2010	TSC	focal
FCD IIa	DEPDC5	c.484‐1G>A	Baulac S, Ishida S, Marsan E, et al Ann Neurol. 2015	FCD	focal
		c.1759C>T (p.Arg587*)	Baulac S, Ishida S, Marsan E, et al Ann Neurol. 2015	FCD	focal
		c.C1663T, p.Arg555*	Scerri T, Riseley JR, Gillies G, et al Annals of clinical and translational neurology. 2015	FCD	hemispheric and quadrantic
		c.856C>T/p.Arg286*	Ribierre, T., et al J Clin Invest. 2018	FCD	focal
	NPRL3	c.1375_1376dupAC (p.S460Pfs*20)	Sim JC, Scerri T, Fanjul‐Fernández M, et al Ann Neurol. 2016	FCD	quadrantic
		c.1352‐4delACAGinsTGACCCATCC	Sim JC, Scerri T, Fanjul‐Fernández M, et al Ann Neurol. 2016	FCD	lobar/sublobar
		c.275G>A (p.R92Q)	Sim JC, Scerri T, Fanjul‐Fernández M, et al Ann Neurol. 2016	FCD	focal
		c.1270C>T, p.Arg424*	Weckhuysen S, Marsan E, Lambrecq V, et al Epilepsia. 2016	FCD	focal
	PCDH19	c.696T>A: p.(Asn232Lys)	Kurian, M., Korff, C. M., Ranza, E., et al Dev Med Child Neurol. 2018	FCD	focal
	CNTNAP2	3709delG	Strauss K, Puffenberger E, Huentelman M, et al New Engl J Medicine. 2006	FCD	focal
FCD Ia	NPRL2	c.68_69delCT p.Ile23Aspfs*6	Weckhuysen S, Marsan E, Lambrecq V, et al Epilepsia. 2016	FCD	focal
	COL4A1	c.1121‐2dupA (for the described histopath)	Yoneda Y, Haginoya K, Kato M, et al Ann Neurol. 2013 65	porencephaly	unilateral
FCD I	DEPDC5	c.1264C>T (p.Arg422*)	Baulac S, Ishida S, Marsan E, et al Ann Neurol. 2015	FCD	focal
	PCDH19	duplication of exons 3,4,5	Kurian, M., Korff, C. M., Ranza, E., et al Dev Med Child Neurol. 2018	FCD	focal

It is notable that the genetic abnormalities in the different types of FCD are similar or occur in similar pathways irrespective of the histological type of FCD. This may suggest a shared pathogenesis between different types of FCD. It does however raise the question of what additional factors, possibly epigenetic, modify the phenotype to determine the histopathological subtype. It will also be important to determine how the relative contribution of genetic, epigenetic and histological features predicts the clinical manifestation and outcome of these patients. Careful interpretation and reporting of genetic findings, both somatic and germline variants, is warranted, as recommended by ACMG and AMP guidelines 39, 57.

Multilobar to hemispheric malformations with somatic variants

Extensive malformations, most typically hemimegalencephaly and megalencephaly, contribute to the clinical picture of selected complex syndromes, such as CLOVES (congenital lipomatous overgrowth, vascular malformations and epidermal naevi), MCAP (megalencephaly‐capillary malformation‐polymicrogyria syndrome), MPPH (megalencephaly‐polymicrogyria‐polydactyly‐hydrocephalus syndrome 1) and Proteus syndrome, and their genetic background is known. Mostly, the involved genes are located upstream in the PI3K/AKT pathway, e.g. PIK3CA, PIK3R, AKT1, AKT3 and others, with mutations being present either in germline or in selected tissues only (brain, skin, blood vessels, etc.). Physical findings consistent with the clinical picture of these syndromes and/or detection of somatic variants in brain tissue samples should prompt the clinician to recommend further genetic consultation. Indeed, purely somatic variants of PIK3CA and AKT3 genes have been detected in the affected brain tissue of patients with hemimegalencephaly 24, 35, 54 with histopathological features of FCD IIa. Interestingly, Lee et al and Leventer discovered variants in MTOR also in hemimegalencephaly and hemispheric cortical dysplasia, respectively, with features consistent with FCD IIa 35, 38.

Mirzaa et al observed even more remarkable findings when they detected a somatic variant of DEPDC5 also in a hemimegalencephaly tissue sample 49, to our knowledge the first case associating somatic mutation in DEPDC5 with hemimegalencephaly [for germline see next paragraph or 59].

We may therefore conclude that at least in certain cases both focal and hemispheric cortical dysplasias share a common genetic background, mostly converging in selected genes of PI3K/AKT/mTOR pathway. The striking question however remains, why in certain cases the entire hemisphere is involved, and in others only a small focal malformation develops. The obvious explanation might be that a mutational event simply occurred later in the development, therefore affecting only a smaller cell population, leading to a more focal brain involvement. In a most recent publication by D'Gama et al, they found that indeed the two‐hit mechanism observed in tumor pathogenesis may be involved in the formation of hemimegalencephaly and FCD. The authors suggest that these two may represent a continuum of pathologies, rather than distinct entities 11. In the case of hemimegalencephaly, however, further studies are needed to confirm whether the opposite hemisphere truly remains unaffected. The evidence of successful epilepsy surgery in patients who achieve freedom from seizures after resection or disconnection of the affected hemisphere in more than 80% of the cases in the recent years 20 would suggest so. On the contrary, an earlier report 21 showed irregular cortical lamination, along with scattered neurons in white matter of the macroscopically unaffected hemisphere in a child with hemimegalencephaly who died of surgical complications after hemispherectomy. Therefore, further molecular genetic studies, especially in postmortem cases, along with studies on animal models 16 seem necessary to elucidate these uncertainties.

Selected focal and hemispheric pathologies with germline variants

As mentioned in the previous paragraph, certain, mostly hemispheric, brain malformations, potentially amenable by epilepsy surgery, may originate as a result of either germline or somatic mutations. In cases where one suspects the involvement of germline variants (Table 2), further genetic testing of these genes may be warranted.

The most obvious case is the tuberous sclerosis complex (TSC), where genetic studies on the presence of germline mutations of TSC1 or TSC2 genes are recommended, although a combination of clinical diagnostic criteria suffices for the definite diagnosis of TSC 53. Despite multiple efforts, the underlying mechanisms of tuber formation have not been clearly explained; second‐hit events expected to lead to tuber formation occur rather less frequently than was expected 47, 55. From recent reports, the role of differential expression of various microRNA types in tuber tissue begins to emerge 61.

More extensive molecular genetic studies comparing tuber tissue, peri‐tuberal dysplastic tissue and seemingly “healthy” brain tissue might bring more light into the genetic and potentially epigenetic factors 27 that distinguish these three “types” of brain changes in TSC patients. In addition, by comparing genetic variants present in these tissue “types” (tuber vs. peri‐tuberal dysplastic tissue vs. macroscopically and microscopically unaffected tissue), we might be able to infer more on the functional effect of such variants.

Except for the aforementioned cases, for a long time, FCD has been considered a solely sporadic disease with no genetic component identified. As described above, somatic variants in multiple genes of the mTOR pathway have been identified to drive the dysplastic changes observed in FCD. However, up until very recently, no one has considered FCD to be a heritable condition. This changed with the publication of the first reports of families with multiple family members affected with FCD 37, familial focal epilepsy, non‐lesional focal epilepsy and their comorbidities with a common genetic origin.

Strauss et al 63 published an interesting report of homozygous mutation of CNTNAP2 in two Old Order Amish siblings with focal epilepsy and cortical dysplasia undergoing epilepsy surgery. Recently, it has been shown that PCDH19, a gene associated with early infantile epileptic encephalopathy resembling Dravet syndrome, may also be involved in the formation of FCD; five patients with variants in PCDH19 had either MRI or MRI and histopathological evidence of FCD 32. Weckhuysen et al 64 reported extensively on families with involvement of all three genes of GATOR1 complex that negatively regulates MTOR complex 1, namely DEPDC5, NPRL2 and NPRL3; the phenotypic spectrum included patients with focal epilepsy, FCD and even SUDEP (sudden unexpected death in epilepsy). Scheffer et al in their study showed a similar phenotypic pattern of FCD, non‐lesional focal epilepsies of various origins with psychiatric co‐morbidities in three families with FCD, with detected germline mutations of DEPDC5 60. Baulac et al and Ribierre et al 5, 56 showed that a germline and a brain somatic mutation in DEPDC5 can co‐occur, as was the case of a single patient from a family affected with FCD type I and focal epilepsy; unfortunately, the tissue specimen was fragmented which might have precluded an unequivocal histopathological diagnosis. In the Ribierre's report, the patient suffered from FCD IIa. Interestingly, Scerri et al 59 even further expanded the spectrum of DEPDC5‐associated phenotype with a report on siblings, one of whom suffered from hemispheric FCD IIa. Sim et al 62 characterized another set of siblings affected with FCD IIa and focal epilepsy and a heterozygous germline mutation in NPRL3, another component of GATOR1 complex. Given that DEPDC5 is known to be a common cause of various familial focal epilepsies 4, e.g. familial focal epilepsy with variable foci (FFEVF), sleep‐related hypermotor epilepsy (SHE), previously known as autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) and temporal lobe epilepsy (TLE), and mutations in NPRL3 were detected in multiple cases of familial focal epilepsies 58, further studies of the role of DEPDC5 and other components of GATOR1 complex in both lesional and non‐lesional epilepsy are necessary. The pathogenic variants of genes of the GATOR1 complex seem to be involved in a broad spectrum of epilepsies ranging from “purely” genetic non‐lesional to “purely” structural based on FCD 2; therefore, they may become ideal candidates for studies of epileptogenesis in both “idiopathic” and structural cases.

Genetic background as a contributing factor in epilepsy surgery outcomes and patient prognosis?

Epilepsy surgery has been established as a safe and effective treatment of drug‐resistant epilepsy in multiple genetic conditions. Leventer et al 37 published evidence on 12 patients from six families with FCD, hemimegalencephaly and DNET; 11 of them underwent epilepsy surgery. Among these, eight patients reached freedom from seizures, two patients had >50% improvement and a single patient suffered ongoing seizures. Unfortunately, no molecular genetic analysis was performed to ascertain the genes possibly involved.

In contrast, genetic analysis is being widely performed in patients with TSC who have increasingly become epilepsy surgery candidates in part due to improved surgical techniques capable to manage such complex cases. Original reports and literature reviews published in recent years state that around 53%–57% of the TSC patients undergoing epilepsy surgery reach freedom from seizures classified as Engel I outcome 25, 28, 45. The factors associated with favorable surgical outcome vary widely; from variables associated with epilepsy itself, e.g. the presence of tonic seizures 25, infantile spasms, younger age at epilepsy onset 45, to EEG patterns, e.g. concordance of ictal and interictal EEG findings and surgery‐related factors such as complete removal of epileptogenic zone, or the need for invasive EEG recording 28. However, none of the authors have established the prognostic value of mutation type (TSC1 vs. TSC2 affected) or specific histopathological changes observed in resected tissue for the prospect of eventual freedom from seizures.

Compared to the number of reports on epilepsy surgery outcomes in patients with TSC, the published outcomes of patients with other clearly genetic conditions is rather scarce. Barba et al identified 12 patients with neurofibromatosis from 25 European epilepsy surgery centers who underwent epilepsy surgery for a broad spectrum of pathologies, including DNET, hippocampal sclerosis and polymicrogyria with 8 patients achieving freedom from seizures one year after surgery 3. However, it is unclear how the disease itself relates to the formation of such a broad spectrum of brain pathologies.

Recently, reports on multiple families with GATOR1‐related FCD have appeared, along with yet limited data on epilepsy surgery in such cases 2. Weckhuysen et al reported two patients with NPRL2 and NPRL3 mutations who achieved seizure reduction but not complete seizure freedom 64. Baulac et al in their article on DEPDC5 mutations in FCD mention five epilepsy surgery patients; three achieved complete seizure freedom, one had worthwhile and one no improvement in seizure frequency 5. Variants in NPRL3 associated with FCDIIa were observed in four patients from three unrelated families undergoing epilepsy surgery in a report by Sim et al 62; three of them became seizure‐free. Some controversy on the pathogenicity of some, particularly missense, variants arises as the reported variants were observed in seemingly unaffected relatives; this could reflect the phenomenon of reduced penetrance in autosomal‐dominant diseases. However, further molecular analyses and functional studies are needed to unequivocally establish the pathogenicity in such cases, as functional studies are considered “strong” evidence for pathogenicity according to the ACMG guidelines 57.

These observations stress the missing connection between how the underlying neurobiology of the disease, the genetic pathways involved reflected in complex cellular and tissue changes influence eventual outcomes of patients undergoing epilepsy surgery for structural genetic epilepsy. We suggest that by incorporating genetic findings, particularly the knowledge of both germline and MCD tissue‐specific variants in selected genes, together with the histological phenotype in the ongoing longitudinal studies on epilepsy surgery outcomes, we may be able to better understand the effect of molecular genetic background of specific MCD on the prognosis of future patients undergoing epilepsy surgery. In future, patients with selected variants in mTOR pathway genes may be eligible to targeted treatment, e.g. mTOR inhibitors, and this would further change the entire landscape of treatment of focal MCD‐related epilepsy; this underscores the importance of molecular genetic diagnosis in MCD patients.

Conclusion and Suggestions for Clinical Practice and Future Research

The aim of this review is to correlate the genetic causes with the underlying morphological abnormalities in focal cortical malformations and to stress the importance of the underlying biology of focal epilepsies of structural genetic origin for patient management and counseling. From the summarized findings of multiple authors, it is obvious that MCD may have a diverse genetic background despite a similar or even identical histopathologic picture. Even though most of their molecular genetic findings converge on various levels of the PI3K/AKT/mTOR pathway, the exact mechanisms underlying MCD formation have not yet been completely described or indeed how this pathway generates a diverse range of histological abnormalities.

The observation that there is a poor correlation between the genetic abnormalities and phenotype in MCD suggests that there are additional genetic or epigenetic changes that are governing the phenotypic manifestation and that a complete description of the underlying biology should include both genetic and pathological data. For both research and clinical purposes, with the addition of genetic diagnosis to our current histopathological diagnosis, we will be able to provide a multilayered diagnosis for MCD patients, as is the current recommended practice for brain tumor diagnostic work‐up (the WHO/Haarlem criteria) 43, 44. For MCD patients undergoing epilepsy surgery, we therefore propose a diagnostic work‐up that would incorporate both somatic and germline DNA testing (Figure 1). Depending on the standard procedures in a given laboratory, quality control report of NGS test and its validation, Sanger sequencing may be required to validate the germline variants. With a multilayered molecular and neuropathological diagnosis and detailed clinical follow‐up information in the post‐surgical period in hand, we could correlate clinical, neuropathological and molecular genetic data in large patient cohorts to draw a complete natural history of selected MCD. In addition, further genetic studies performed on non‐diseased tissue controls might shed new light on the prevalence of variants of uncertain significance in seemingly unaffected brain tissue; currently, to our knowledge, no studies comparing variants present in control non‐disease tissue samples vs. MCD samples have been performed. This fact only emphasizes the importance of incorporating molecular genetic methods in clinical neuropathological practice, as more data would be obtained to elucidate the actual pathogenic effect of such variants. Given the complexity of molecular genetic testing and results’ interpretation for MCD cases, we suggest a two‐tiered approach toward incorporating molecular genetic methods in clinical and research practice (see Figure 2). In the first step, centers for epilepsy and epilepsy surgery might adopt targeted gene panel testing of genes with known germline variants associated with MCD. Gene panel testing for germline variants has been applied in clinical practice for many conditions, including epilepsies 36, hereditary neuropathies 34 and others, and it would represent a valuable diagnostic tool to add to our current diagnostic process for patients with focal drug‐resistant epilepsy caused by MCD. In the second step, the advanced centers for epilepsy surgery, along with the necessary collaboration of associated university research teams, may employ advanced molecular genetic methods to detect rare somatic variants in brain tissue samples. Such methods could be based on, e.g., targeted ultra‐deep sequencing of a selected set of genes involved in MCD formation. This review aims to provide a summary of methods and approaches (see Supplementary materials) that have been employed by various research centers and provide theoretical background for centers that intend to adopt advanced multilayered diagnostic processes for MCD patients.

Figure 1.

Overview of the proposed diagnostic work‐up for patients with MCD.

Figure 2.

Overview of the proposed approach toward incorporating molecular genetic methods in clinical and research practice. Brain by Simon Stratford from the Noun Project. Gene by Nithinan Tatah from the Noun Project. Human body by H Alberto Gongora from the Noun Project.

Such information obtained from molecular genetic studies, combined with detailed information on patients’ clinical presentation would enable us to provide patients and their neurologists with a more accurate prognosis for seizure freedom that might guide postoperative patient management (e.g. antiepileptic drugs withdrawal, re‐operation planning, family counseling, etc.). While the prognostic benefit of such a system remains unproven, future studies that stratify patients based on both histological and genetic diagnosis would allow a more comprehensive analysis of the underlying disease in these patients and would allow us to determine the key biological events that correlate with the clinical outcome. Based on our findings (see Figure 3), we therefore propose that in future, patients diagnosed and operated for drug‐resistant epilepsy should have an integrated molecular and pathological diagnosis similar to the current practice in brain tumor diagnostic processes (see Figure 1).

Figure 3.

Summary of the genes involved in formation of FCD type IIa, type IIb and type I. The bottom photomicrograph shows FCD type Ia, but the genes described are involved in the formation of the respective types of FCD type I (for details see Tables 1 and 2). The schematic picture of the brain refers to somatic variants and the schematic picture of the human body refers to germline variants. Brain by Simon Stratford from the Noun Project. Human body by H Alberto Gongora from the Noun Project.

Acknowledgments and Funding

We are grateful for funding from the Brain Tumour Charity, Great Ormond Street Hospital Children's Charity, Children with Cancer UK, Cancer Research UK, the Olivia Hodson Cancer Fund and the National Institute for Health Research. The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health. The authors have no conflict of interest to declare. We would like to thank Professor Josef Zamecnik, M.D., Ph.D. from Motol University Hospital, Prague, Czech Republic and Dr. Yao‐Feng Li from UCL Great Ormond Street Institute of Child Health, London, UK for providing the photomicrographs.

Supporting information

Table S1. Summary of molecular genetic and histopathological findings from the reviewed literature, along with methods used in the reports.