Malformations of Cortical and Brain Development – General Approach

A heterogeneous class of disorders of cerebral cortex formation (proliferation, migration and post-migrational development), of germline genetic, somatic mosaic or acquired cause, with in-depth coverage of polymicrogyria, schizencephaly, focal cortical dysplasias of the mTOR pathway and the GATOR1 complex, megalencephalies and hemimegalencephaly of the PI3K–AKT–mTOR pathway, heterotopias and SLC35A2-associated MOGHE.

Literature and regulatory information update: 26 October 2026.

0. Initial Reference Table

Entities in scope and related entities

Related conditions: primary microcephaly with MCD, agenesis of the corpus callosum, posterior fossa and Dandy-Walker malformations, pontocerebellar hypoplasias, Joubert syndrome, Zellweger spectrum, Aicardi-Goutières syndrome and TSC [5, 7, 13, 16, 23, 30, 35].

1. Introduction

1.1. Historical overview

Malformations of cortical development (MCD) are a large and heterogeneous group of disorders of cerebral cortex formation, caused by genetic, infectious, vascular or metabolic factors, which result in an abnormal cortical structure or heterotopic grey matter, sometimes accompanied by abnormal brain size (micrencephaly or megalencephaly) [1]. They are a major cause of epilepsy, developmental delay, intellectual disability (ID), autism spectrum disorder (ASD) and cerebral palsy, with symptoms appearing mainly between early childhood and early adulthood; an estimated 40–50% of drug-resistant epilepsies treated surgically in children are due to an MCD [1, 5].

The history of focal dysplasias begins in the neuropathology of epilepsy surgery. In 1957, Crome described a form of “ulegyria” with irregular nerve cells, and in 1971 David Taylor and colleagues coined the term “focal cortical dysplasia” (FCD) from the lobectomy specimens of 10 patients with drug-resistant epilepsy: the abnormality was rarely identifiable by palpation or the naked eye, and microscopy showed clusters of large, bizarre neurons in all layers except the first and, in most cases, grotesque cells, probably of glial origin, in the deep zones of the cortex and the underlying white matter; the authors noted the resemblance to tuberous sclerosis but considered that the clinical and pathological differences justified a separate entity [7, 36]. These were followed by the classifications of Mischel (1995, 77 specimens) and Palmini (2004), who divided FCD into types I and II with two subtypes each, and the first consensus classification of the International League Against Epilepsy (ILAE), in 2011, which subdivided type I into three subtypes, kept Palmini's type II and introduced type III, associated with another principal lesion [6, 7].

Polymicrogyria (PMG) entered clinical practice through congenital bilateral perisylvian syndrome, described in a multicentre study of 31 patients published in 1993: pseudobulbar palsy with facial, pharyngeal and masticatory diplegia, cognitive impairment and epilepsy, with bilateral perisylvian PMG on MRI confirmed at autopsy [37]. In 2004, Piao and colleagues identified variants in GPR56 (now ADGRG1) as the cause of so-called “bilateral frontoparietal PMG”, the first G protein-coupled receptor gene implicated in regional development of the human cortex [38].

The years 2012 to 2015 brought the molecular explanation of overgrowth malformations and of type II FCD. In 2012, three studies showed de novo germline or post-zygotic variants in AKT3, PIK3R2 and PIK3CA in the MPPH and MCAP megalencephaly syndromes, and somatic variants in PIK3CA, AKT3 and MTOR, or mosaic trisomy 1q, in resected hemimegalencephaly tissue [39–41]. In 2013, exome sequencing linked the DEPDC5 gene, a component of the GATOR1 complex, to familial focal epilepsies [42, 43], and in 2015 brain somatic gain-of-function variants in MTOR were described in 15.6% of a cohort with type II FCD [44]. In 2017 MOGHE was recognised, a white matter neuropathological entity with oligodendroglial hyperplasia, soon associated with brain mosaicism for SLC35A2 [7, 35].

1.2. Current classifications

Barkovich's developmental and genetic classification, successively updated (1996, 2001, 2005 and 2012), organises MCD by the stage of cortical development primarily affected, recognising that an early alteration disrupts the subsequent stages and that the processes run simultaneously in different regions [1, 4]. In the 2012 update, which brings together about 200 clinical entities, Group III changed from “abnormal cortical organisation” to “abnormal post-migrational development”, because organisation begins before migration ends, and Group IV (“unclassified”) was eliminated; entities were ordered by mode of inheritance and by clinical or genetic definition, and the authors acknowledged that this was an intermediate system, to be replaced by a classification based on biological pathways [4, 5].

The table summarises the structure described in the main text of the 2012 update and in the 2020 consensus; the complete lists of entities and genes are in the appendix and supplementary tables of the original paper [1, 4]. The European Neuro-MIG consensus (COST network CA16118) added practical imaging definitions and stressed that the classifications are “fluid” models: different variants in the same gene produce different MCDs, different genes in the same pathway produce the same pattern, and distinct mechanisms converge on the same imaging phenotype; PMG remains classified as post-migrational, but histopathological evidence of pial limiting membrane defects and leptomeningeal heterotopia suggests that it is a late migration anomaly [1]. The Neuro-MIG recommendations resulted from systematic reviews (about 500 papers between January 2018 and September 2019) and consensus meetings, including one in Lisbon in 2018, without formal grading of evidence owing to the lack of randomised trials [1, 5].

For FCD, the 2011 ILAE classification was updated in 2022 by a Task Force that reviewed 1349 publications, surveyed 367 professionals (75.1% used the ILAE classification, more than 60% asked for genetics to be incorporated and more than 50% for mMCD to be included) and tested agreement among 20 neuropathologists from 16 countries: interobserver agreement rose to kappa 0.65 when genetic results were made available [7]. The update kept types I, II and III, added two white matter categories (mMCD and MOGHE) and a descriptive category (“no definite FCD”), and proposed an integrated diagnosis in three layers (histopathology, genetics and neuroimaging) [7].

In the absence of the recommended immunohistochemistry (NeuN, non-phosphorylated neurofilaments, vimentin, MAP2, CD34, OLIG2, GFAP, αB-crystallin), the type should be recorded as “not otherwise specified” (NOS) [7]. The lissencephaly spectrum follows the Di Donato imaging classification (2017), which defines 21 recurrent patterns by the anteroposterior gradient, the degree of severity, the thickness and appearance of the cortex and the associated non-cortical malformations, and predicts the likely gene; the old division into type I and type II lissencephaly has been abandoned, and cobblestone malformation no longer belongs to the group [1, 5, 8]. Subcortical heterotopias received their first imaging classification in 2019, with five groups [5]. PMG is described by topography (seven patterns in the 2018 GeneReview; six topographic patterns and 13 morphological subtypes in the 2020 consensus) and by at least six syndromes defined by radiological and clinical criteria [5, 13].

1.3. Definitions of the entities in scope

The consensus definitions, with Human Phenotype Ontology (HPO) identifiers, are those of Oegema and colleagues [5]. The imaging consensus defines microcephaly as HC below −2 SD (3rd centile) for age and sex, mild between −2 and −3 SD (0.3rd centile), and macrocephaly as HC above +2 SD (97th centile), mild between +2 and +3 SD; the Oxford Desk Reference uses the +3 SD threshold for macrocephaly [1, 45]. Megalencephaly designates an enlarged brain and is distinguished from macrocephaly due to other causes (hydrocephalus, enlarged extra-axial spaces, skeletal dysplasias); it may be bilateral or unilateral, complete or partial, and classic HME corresponds to complete megalencephaly of one hemisphere, whereas forms involving up to three lobes (“quadrantic dysplasia”, lobar HME) result mostly from somatic mosaic variants [1].

FCD is a spectrum of focal malformations with disordered cortical lamination, with or without abnormal cell types; type II and dysplastic megalencephaly, including classic HME, form a continuum caused by germline, somatic or “two-hit” (germline and somatic) variants in mTOR pathway genes, and TSC tubers are considered a subtype of FCD IIb, distinguished by the TSC criteria [1, 5]. Bottom-of-sulcus (BOS) FCD is a lesion restricted to the bottom of a sulcus, mainly frontal, which on imaging anticipates an FCD type II (usually IIb) [7]. mMCD is characterised by an excess of heterotopic neurons in the white matter (more than 30/mm²) without another principal lesion, and MOGHE combines heterotopic neurons in the white matter with an oligodendroglial density above 2200 OLIG2-positive cells/mm² (normal below 1000/mm²) [7].

PMG is a malformation with abnormal cortical lamination and an excessive number of small gyri, which may be focal, multifocal or generalised, unilateral or bilateral, symmetric or asymmetric [1, 13]. Schizencephaly is a full-thickness cleft lined by PMG or heterotopia, from the ependyma to the pia (pial–ependymal seam), with open lips (patent cleft with cerebrospinal fluid, CSF) or closed lips (apposed margins, with a “dimple” at the ventricular margin), and is distinguished from porencephaly, a clastic lesion lined by white matter with or without gliosis [1]. Heterotopias are clusters of normal neurons in an abnormal location, isointense to cortex on all sequences: periventricular nodular heterotopia, the most frequent; subcortical heterotopia (usually transmantle, from the ventricle to the cortex); subcortical band heterotopia (formerly “double cortex”), thick (more than 7 mm) or thin (1–7 mm); and ribbon-like heterotopia, with bilateral giant undulating bands [1]. Dysgyria designates an abnormal gyral pattern with normal cortical thickness, first described in tubulinopathies and dystroglycanopathies [1, 26].

1.4. Scope and organisation of the article

This article covers the MCD class uniformly, with the common syndromic description, the classifications and the table of types and genes, and goes into depth on the following entities: PMG, schizencephaly, FCD and mMCD, MOGHE, megalencephalies and HME, MPPH, SKS, GATOR1 epilepsies and subcortical heterotopias. The lissencephaly spectrum and Miller-Dieker syndromes, the cobblestone malformations of the dystroglycanopathies, FLNA-related periventricular nodular heterotopia, pontocerebellar hypoplasias, holoprosencephaly, primary microcephaly, agenesis of the corpus callosum, TSC, the PIK3CA-associated PROS spectrum, Joubert syndrome, Zellweger spectrum peroxisomal disorders and Aicardi-Goutières syndrome are summarised where they bear on the general approach. The PMG GeneReview was withdrawn in 2024 because the phenotype was too broad; data from that archived version (2018) are used as historical and updated with the 2020 and 2022 consensus statements [1, 5, 7, 13].

2. Molecular Genetics and Pathophysiology

2.1. Timeline of cortical development and general mechanisms

Neuronal proliferation in the dorsal telencephalon begins at 5–6 weeks of gestation, declines after 16 weeks and ends at 22–25 weeks; migration also begins at 5–6 weeks, peaks between the third and fifth months and ends at 30–35 weeks, while cerebellar migration continues until the middle of the second year of life; post-migrational development (axonogenesis, dendritogenesis, synaptogenesis and pruning) extends beyond the intrauterine period and some processes into adulthood [1]. Apical radial glia serve as the scaffold for radial migration, and outer radial glia of the outer subventricular zone are decisive for the expansion and gyrification of the human cortex; inhibitory interneurons arise in the medial and caudal ganglionic eminences and reach the cortex by tangential migration [1]. On fetal ultrasound the brain outline is smooth until about 24 weeks, from which point the first sulci and gyri appear [29].

Each stage has its own disease mechanisms. Reduced proliferation or increased apoptosis gives rise to microcephaly, increased proliferation or reduced apoptosis to megalencephaly, and abnormal proliferation (dysgenesis) produces the dysmorphic neurons of FCD type II and of megalencephaly with cortical malformation, now attributed to germline and somatic activating variants of the mTOR pathway in progenitors of excitatory neurons of the dorsal telencephalon [1]. In migration, disruption of the neuroependyma by abnormal vesicle trafficking leaves next to the ventricle the neurons that have lost their connection to the radial glia (periventricular nodular heterotopia); disturbance of the molecular machinery of migration (tubulins, microtubule-associated proteins such as LIS1 and DCX, actins) produces lissencephaly, microcephaly with lissencephaly or dysgyria; and defects of the pial surface allow neurons to cross the basement membrane into the leptomeninges, with the PMG–cobblestone spectrum, which can result from dystroglycanopathies, lamininopathies or ADGRG1 variants as well as from ischaemic, teratogenic or infectious injury; the working hypothesis is that small pial breaches produce PMG and large breaches produce cobblestone malformation [1]. Isolated post-migrational disorders include secondary microcephaly, FCD type I and certain dysgyrias [1].

Inborn errors of metabolism fit into this timeline: disorders affecting morphogenesis (first 6–8 weeks) produce malformations, and those affecting growth and differentiation (after 8–9 weeks) produce dysplasias, a category that includes metabolic brain malformations; those causing irreversible anomalies of the central nervous system (CNS) include disorders of cellular trafficking, O-glycosylation, cholesterol synthesis and amino acid synthesis and transport, non-ketotic hyperglycinaemia (NKH) and deficiency of the docosahexaenoic acid transporter MFSD2A [46].

2.2. The PI3K–AKT–mTOR pathway: activators, repressors and mosaicism

The PI3K–AKT–mTOR pathway regulates protein synthesis, metabolism, the cell cycle, survival, and cell growth and proliferation; gain-of-function variants in PIK3CA, PIK3R2, AKT3, CCND2 and MTOR produce a spectrum of diffuse and segmental brain disorders ranging from megalencephaly and HME to PMG and FCD [9]. AKT3 is the main target of phosphatidylinositol-3,4,5-trisphosphate (PIP3), PIK3R2 encodes the regulatory beta subunit (p85-beta) of phosphoinositide 3-kinase (PI3K) and cyclin D2 acts downstream, at the G1–S transition of the cell cycle [9]. The recurrent variant PIK3R2 p.(Gly373Arg), in the inter-SH2 domain, prevents p85-beta from restraining the catalytic subunit p110, with increased PIP3 and AKT and mTOR signalling [16]. The MPPH variants of CCND2 lie in the last exon, around a residue phosphorylated by GSK-3β; mutant cyclin D2 resists proteasomal degradation and accumulates, with excessive proliferation of neural progenitors, and cells with variants in PIK3CA, PIK3R2 or AKT3 show the same accumulation, suggesting a unifying mechanism for the megalencephalies of the pathway [16]. In tissue from patients with megalencephaly, PIP3 and phosphorylation of S6 and 4E-BP1 are increased, functional proof of gain of function [39].

The GATOR1 complex (DEPDC5, NPRL2 and NPRL3) represses mTORC1 in the amino acid-sensing branch: under amino acid depletion it acts as a GTPase-activating protein (GAP) for RagA/RagB, which switch to the inactive GDP-bound form, releasing mTORC1 from the lysosomal surface; DEPDC5 contacts the nucleotide-binding pocket of the Rags, but the GAP reaction is carried out by NPRL2 [11, 16]. DEPDC5 is a 1603-amino-acid protein encoded at 22q12.2-q12.3, and the mechanism of pathogenic variants is loss of function, with several nonsense variants degraded by nonsense-mediated decay (NMD); in a functional study, two nonsense variants and a 3 bp deletion disrupted mTORC1 inhibition, but none of seven missense variants did, which calls for caution in interpreting missense variants [11, 16]. NPRL3, at 16p13.3, overlaps the regulatory region of the alpha-globin gene cluster, so a whole-gene deletion can give a dual diagnosis of NPRL3 epilepsy and alpha-thalassaemia trait [16].

The difference between activators and repressors determines the number of hits required. A single gain-of-function hit in an activator (PIK3CA, AKT3, RHEB) or in MTOR is enough to produce FCD type II, whereas in the repressors (DEPDC5, NPRL2, NPRL3, TSC1, TSC2) a double loss-of-function hit, germline and somatic, is required; biallelic somatic second hits in DEPDC5 (point variant or copy-neutral loss of heterozygosity) have already been demonstrated [7, 11, 47]. In mosaic human cortical organoids, biallelic inactivation of DEPDC5 reproduced increased mTOR activity reversed by rapamycin, dysmorphic-type neurons and hyperexcitability, reinforcing the need for mosaic biallelic inactivation (preclinical study) [48].

The timing and the progenitor in which the variant arises determine the extent of the lesion. Deep sequencing of pathway genes in FCD and HME tissue established the aetiology in 27 of 66 cases (41%), with radiologically indistinguishable lesions caused by somatic activating variants in AKT3, MTOR and PIK3CA and by germline loss-of-function variants in DEPDC5, NPRL2 and TSC1/TSC2; variants in the same gene cause a continuum FCD → HME → bilateral brain overgrowth, depending on the developmental timing, and in mice conditional activation of Pik3ca in excitatory neurons and glia, but not in interneurons, was sufficient for cortical overgrowth [49]. The mosaic fraction is greater in HME (an earlier event), the allelic fraction gradient follows the gradient of dysmorphic neuron density (maximal in the seizure onset zone), microdissected dysmorphic neurons and balloon cells contain the variant, and synergy of two mosaic variants (RPS6 and MTOR) has been described in one HME [7]. Some dysplastic HMEs correspond to large areas of FCD type II, and HME may extend to non-cerebral tissues, with skin changes and localised overgrowth [1, 5].

2.3. mTOR pathway genes and associated entities

MTOR

MTOR (1p36.22) encodes the serine/threonine kinase that forms the catalytic core of the mTORC1 and mTORC2 complexes; all causal variants described are activating missense variants in the coding region, clustered in the kinase and FAT domains, without intragenic deletions or duplications, and loss of function is not a described mechanism [10, 16]. There are two distinct situations: constitutional or early post-zygotic variants (20–25% mosaic) cause SKS, whereas somatic variants restricted to the brain cause apparently isolated FCD type II (OMIM 607341) and HME with refractory epilepsy [10, 16]. The variant NM_004958.4:c.5395G>A p.(Glu1799Lys) accounts for almost 50% of SKS cases [10, 14]. In FCD type II, deep exome sequencing of brain–blood pairs identified somatic NM_004958.4:c.7280T>C p.(Leu2427Pro) in 2 of 4 patients, and targeted sequencing of a further 73 patients found eight different somatic variants in 10, a total of 12/77 (15.6%); in utero electroporation of mutant MTOR in mice disrupted neuronal migration and produced cytomegalic neurons and spontaneous seizures, suppressed by rapamycin [44]. The OMIM entry for FCD type II (FCORD2, 607341) brings together three genes with somatic variants: MTOR (1p36.22), TSC1 (9q34.13) and TSC2 (16p13.3, with the relationship flagged as provisional) [15]. The series summarised in OMIM confirm low allelic fractions in tissue: in the 2015 study, brain MTOR variants (about 1–12%) were absent from blood; in another 2015 series, somatic MTOR variants were found in 6 of 13 FCD IIb (46%), with fractions of 1.11–9.31%, and increased S6K phosphorylation in lesional tissue; in 2016, in 6 of 16 FCD type II (37%), with fractions below 7% and intense S6 phosphorylation in dysmorphic neurons but not in adjacent normal neurons; and in 2017 de novo somatic variants in TSC1 (4 patients) and TSC2 (1 patient) were identified in 5 of 40 FCD type II without an MTOR variant (12.5%), with fractions below 3%, without clinical signs of TSC [15]. In cells transfected with these variants, abnormal S6K phosphorylation was inhibited by rapamycin, and in mice inactivation of Tsc1 or Tsc2 by in utero genome editing produced neurons resembling those of FCD type II and seizures, almost completely reversed by rapamycin in the Tsc2 model (preclinical studies) [15]. MTOR is the most frequent causal gene of FCD IIa and IIb and the second most frequent of HME [32].

AKT3, PIK3R2, CCND2 and PIK3CA: the megalencephalies

MPPH results from heterozygous variants in AKT3 (1q43-q44; 30% of the 41 patients with molecular confirmation), CCND2 (12p13.32; 30%) or PIK3R2 (19p13.11; 40%), mostly de novo germline, with somatic mosaicism in some, especially in PIK3R2 and AKT3 [9]. Half of the AKT3 diagnoses correspond to 1q43-q44 duplications detected by CMA, and somatic duplication of the same locus has been found in HME and FCD [9, 41]. In the 2012 study, variants in AKT3, PIK3R2 or PIK3CA explained 37 of 50 unrelated probands (74%): recurrent PIK3R2 p.(Gly373Arg) in 11 MPPH families, de novo in all seven with parental DNA, including the first familial recurrence (three siblings) due to germline mosaicism in a parent; 15 PIK3CA variants, mostly post-zygotic, in 23 patients with MCAP and 1 with MPPH, with mutant levels of 10–50% and, in some, of 1–8% detected only by ultra-deep sequencing (more than 10,000×); and two de novo variants in AKT3, p.(Arg465Trp) and p.(Asn229Ser) [39]. The gene–phenotype correlation was strong: PIK3R2 was associated with MPPH (p = 3.3 × 10⁻⁶) and PIK3CA with MCAP (p = 1.0 × 10⁻⁶) [39].

PIK3R2 is the most frequent identified cause of isolated bilateral perisylvian PMG: in screening of 118 children with this malformation, 20 had PIK3R2 variants (7 with isolated PMG, 13 with MPPH features), 19 with p.(Gly373Arg) and 1 with p.(Lys376Glu), 12 constitutional and 8 mosaic; mosaicism ranged from 3% to 37% of reads (5–73% of cells), and in blood from undetectable to 17%, versus 29–43% in saliva [16, 50]. A PIK3R2 variant therefore does not imply the full MPPH picture [16].

In HME, the study of 8 resected specimens showed mosaic trisomy 1q in 2 (about 40% of cells, absent from blood) and somatic AKT3 NM_005465.7:c.49G>A p.(Glu17Lys) in 1 of the remaining 6, absent from leucocytes — an exact paralogous variant of AKT1 p.(Glu17Lys) (Proteus syndrome) and AKT2 p.(Glu17Lys) (hypoglycaemia and asymmetric growth); AKT3 is the most abundant paralogue in the brain during neurogenesis [41]. In another series of 20 HMEs with brain–blood pairs, 30% had de novo somatic variants in PIK3CA, AKT3 or MTOR, including the recurrent PIK3CA c.1633G>A p.(Glu545Lys) in 4 cases, present in 8–40% of alleles from several brain regions [40]. Loss-of-function variants in the same genes produce the opposite phenotype: AKT3 deletions and proximal truncating CCND2 variants cause microcephaly and ID [9, 16].

GATOR1 complex: DEPDC5, NPRL2 and NPRL3

Heterozygous germline GATOR1 variants explain up to 37% of familial focal epilepsies and together constitute the group of genes most frequently involved in genetic focal epilepsies, with more than 180 families described [30, 43]. The discovery showed DEPDC5 in two families with FFEVF and in 5 of 6 large published families, and in about 12% (10/82) of small families with focal epilepsy [42]. In the review of 183 probands, 140 distinct variants were distributed among loss-of-function (68% pathogenic), likely pathogenic (14%), variants of uncertain significance (VUS, 15%) and likely benign (3%); DEPDC5 accounted for 85% of cases (155/183), and the only recurrent missense variant was NPRL3 p.(Arg92Gln) [30]. In the cohort of 73 new probands, the variants were inherited in 96% (45/47) and, of these, 64% (29/45) from an asymptomatic parent, reflecting reduced penetrance [30]. Truncating variants are more associated with malformations (FCD and HME), but there is no definitive genotype–phenotype correlation [11, 30]. NPRL2 explains about 1.0% of sporadic and 6.3% of familial cases of sleep-related hypermotor epilepsy (GeneReviews series), and NPRL3 is one of the ten genes of this syndrome [16].

TSC1, TSC2, RHEB, PTEN and others

TSC shares the pathway and the two-hit mechanism: tubers are considered FCD IIb, and a germline TSC2 variant with a somatic second hit also explains apparently isolated FCD IIA and isolated HME [1, 5, 49]. Somatic variants have also been described in RHEB (FCD type II, HME), PTEN (megalencephaly, HME, FCD type II; in the germline form, PTEN hamartoma tumour syndrome, PHTS) and variants in PIK3C2B in focal epilepsies [11]. STRADA (autosomal recessive) causes polyhydramnios, megalencephaly and symptomatic epilepsy syndrome, with early lethality [9].

2.4. SLC35A2, glycosylation and MOGHE

SLC35A2 (Xp11.23) encodes the main UDP-galactose transporter of the Golgi apparatus, required for galactosylation of N-glycans, proteins and sphingolipids; loss of function produces truncated glycans lacking galactose [35]. De novo germline variants cause SLC35A2-CDG (congenital disorder of glycosylation type IIm, OMIM 300896), a developmental and epileptic encephalopathy of neonatal or early infantile onset, with spasms and hypsarrhythmia, severe ID and delayed myelination, with female predominance owing to reduced viability of hemizygotes, and with systemic post-zygotic mosaicism in both sexes [16, 35]. Somatic variants restricted to the brain, absent from blood, are the genetic marker of MOGHE: 9 of 20 (45%) in a German surgical series, with variant allele fractions (VAF) of 7–52%, and 17 of 18 SLC35A2-positive cases from an international cohort reclassified as MOGHE after blinded review with NeuN and OLIG2, with no SLC35A2-negative case receiving this diagnosis [35]. The largest cohort, of 47 patients with brain somatic variants, showed histological MOGHE in 44 and 42 distinct variants (33.3% missense, 30.9% frameshift, 23.8% nonsense, 9.5% in-frame indels, 2.4% splicing), with VAF of 1.4–52.6% (mean 17.3%) [51]. In a meta-analysis of 163 histologically confirmed MOGHE cases with molecular data, about 73% had an SLC35A2 variant, and variants in this gene are found in about one third of lesions previously classified as mMCD or FCD type I, a group mechanistically distinct from the mTOR spectrum [16]. All SLC35A2 variants previously attributed to FCD Ia have been reassigned to MOGHE [7].

Microdissection in a male patient (an X-linked gene, so the VAF corresponds to the proportion of cells) showed the variant in 36.5% of glial cells, 28.6% of heterotopic neurons, 8.7% of normal cortical neurons and 50.5% of OLIG2-positive cells in the clusters, indicating origin in a neuroglial progenitor, probably radial glia, before differentiation [35]. The 2026 review describes, in models of MOGHE and of mTORopathies, a convergence towards reduced action potential firing [52].

2.5. Genetics of polymicrogyria

PMG is aetiologically heterogeneous: it has chromosomal causes (22q11.2 and 1p36 deletions, and more rarely 4q21, 6q26, 13q3, 18p11, 21q2 and 2p13 duplication), monogenic causes (more than 40 genes) and non-genetic causes, mainly congenital CMV infection, intrauterine hypoperfusion and other teratogens such as alcohol [13, 29]. The imaging consensus adds, among the monogenic causes, COL4A1/COL4A2, OCLN, RTTN and GRIN1, and among the non-genetic causes Zika, trauma, ischaemic arterial infarcts, twin–twin transfusion syndrome and death of a monozygotic co-twin, with injury to the distal radial glia, pia or arachnoid demonstrated in fetal pathology [1]. The monogenic genes fall into mTORopathies, tubulinopathies, cobblestone dysplasias (alpha-dystroglycanopathies and others, such as lamininopathies and congenital disorders of glycosylation, CDG) and other pathways [13].

The table condenses the gene table of the archived version of the GeneReview, which does not include metabolic, peroxisomal and mitochondrial disorders; the HC labelling by gene was reconstructed from tabular text and serves as a guide [13]. In MCAP, 21 of 24 patients described had somatic PIK3CA mosaicism and 2 of 24 a de novo germline variant [13]. Outside the table are Aicardi syndrome (almost uniform PMG), Mowat-Wilson, Yunis-Varon, Adams-Oliver and Sturge-Weber, in which PMG is infrequent [13]. The best-known gene for PMG with microcephaly is WDR62, and in unilateral PMG a 2026 series (35 individuals from 30 families) found a probable genetic cause in 26.7% of unrelated individuals, with recessive genes (ASPM, WDR62, TMEM216) and dominant genes (22q deletion, DYNC1H1, SCN3A and the hereditary haemorrhagic telangiectasia genes ACVRL1 and ENG), a yield comparable to that of bilateral PMG [5, 53].

ADGRG1 (GPR56)

GPR56 is expressed in cortical progenitors and migrating neurons, binds collagen III and signals via the G12/G13–RhoA pathway, maintaining the integrity of the pial basement membrane and preventing overmigration; pathogenic missense variants affect trafficking and surface expression or the interaction with collagen III, with rupture of the basement membrane and neuronal herniation [16]. The malformation, described as “bilateral frontoparietal PMG”, is in fact a frontally predominant cobblestone malformation, without ocular or muscle involvement, with cerebellar dysplasia and diffuse myelination anomalies [4, 5, 13]. In one cohort, 29 of 29 patients with the typical pattern were homozygous for GPR56 variants, and none with bilateral frontal, bilateral perisylvian or bilateral generalised PMG; in a series of 30 patients with bifrontoparietal PMG, 14 from 8 consanguineous families were homozygous [16]. A second entity, bilateral perisylvian PMG restricted to the cortex around the Sylvian fissure (OMIM 615752), results from a 15 bp deletion in a regional regulatory element of GPR56, which blocks the binding of RFX factors and abolishes expression only in that territory [16].

Tubulinopathies, dynein and kinesins

Heterozygous missense variants in TUBA1A, TUBB2A, TUBB2B, TUBB3, TUBB and TUBG1, with a probable dominant-negative effect, and biallelic variants in TUBA8 cause an overlapping spectrum of malformations (lissencephaly, microlissencephaly, pachygyria, PMG, dysgyria and simplified gyral pattern) with dysmorphic basal ganglia, corpus callosum anomalies, vermian dysplasia with “diagonal” folia and asymmetric brainstem hypoplasia [5, 26, 29]. TUBB2B was the first tubulin gene linked to PMG, with de novo variants in four individuals and in a 27-week fetus with asymmetric bilateral PMG and ectopic neurons in the leptomeninges through breaches in the pial basement membrane [16]. More than 95% of variants arise de novo [26]. DYNC1H1 and KIF2A cause pachygyria to dysgyria with a large caudate and vermian hypoplasia, and DYNC1H1 is frequently accompanied by peripheral motor neuropathy (from fetal akinesia to lower limb-predominant spinal muscular atrophy) [5, 54]. KIF5C variants cluster at residue Glu237 of the microtubule-binding domain (5 of 6 in one series) [16].

2.6. Genetics of the heterotopias

Periventricular nodular heterotopia has more than 20 genes and many CNVs; the mechanisms involve the radial glial scaffold, cell–cell adhesion, vesicle trafficking, PI3K–AKT–mTOR or SMAD2/3 dysregulation, RNA processing and transcriptional regulation [5]. The classic female form is due to loss of function of FLNA (Xq28), lethal in most male fetuses; about 10% of boys with periventricular heterotopia have FLNA variants with partial loss of function [29]. Periventricular heterotopia with microcephaly is due to ARFGEF2 (autosomal recessive), and biallelic variants in DCHS1 and FAT4, a cadherin receptor–ligand pair, cause a recessive syndrome with periventricular heterotopia [29]. NEDD4L (periventricular nodular heterotopia type 7) associates bilateral heterotopia with perisylvian PMG, cleft palate and 2–3 toe syndactyly, with missense variants almost always in the catalytic HECT domain, which dysregulate AKT–mTOR signalling, and a recurrent variant in the WW domain, NM_001144967.3:c.623G>A p.(Arg208Gln) [14, 16]. The 6q27 microdeletion is associated with posteriorly predominant heterotopia with overlying PMG and posterior fossa anomalies [5].

Bilateral symmetric subcortical heterotopias are rare and usually recessive: the mesial parasagittal form of Chudley-McCullough syndrome (GPSM2), ribbon-like heterotopia with agenesis of the corpus callosum and megalencephaly (EML1, a primary cilium gene) and the peritrigonal form (TUBB, KATNB1, CENPJ); curvilinear transmantle heterotopia has no identified genetic cause, and disruptive vascular causes (twinning, threatened miscarriage, trauma) and post-zygotic mosaicism have been suggested [5, 16]. Subcortical band heterotopia belongs to the lissencephaly spectrum: diffuse SBH with a band thicker than 5 mm is pathognomonic of DCX in females and males, and posterior SBH points to mild or mosaic PAFAH1B1 variants [5].

2.7. Schizencephaly and porencephaly

Schizencephaly probably results from an early prenatal focal injury of the germinal matrix or an infarct in a very immature brain, with liquefaction [1]. A population study from California (about 4 million births, 1984–2001) associated it with young maternal age and monozygotic twin pregnancy; one third of cases had anomalies outside the CNS and more than half were classifiable as vascular disruption (gastroschisis, intestinal atresias, amniotic bands) [4]. The initial association with heterozygous germline EMX2 variants in severely affected patients (1996) was not confirmed: two series of 84 and 39 probands found no pathogenic variants, Barkovich's classification considers the gene highly unlikely as a cause and advises against testing, the 2020 consensus states that evidence of a causal role is lacking, and ClinGen curation classifies the association as “Limited” (2024), with other curations ranging from “Moderate” to “Disputed” [4, 5, 16]. Orphanet lists EMX2 only as a major susceptibility factor for acquired schizencephaly [16].

Collagen IV provides the best-documented monogenic cause. COL4A1 and COL4A2 encode the α1 and α2 chains of collagen IV of vascular basement membranes, and heterozygous variants cause a small vessel disease with haemorrhage and ischaemic injury of prenatal, perinatal or postnatal onset, whose characteristic brain expression is porencephaly [27, 55, 56]. In screening of 61 patients with porencephaly and 10 with schizencephaly, 15 (21%) had COL4A1 variants, 10 with porencephaly and 5 with schizencephaly, with intracranial calcification, FCD, pontocerebellar atrophy, ocular changes, myopathy, raised creatine kinase (CK) and haemolytic anaemia as associated findings; these were the first confirmed associations of COL4A1 with schizencephaly and haemolytic anaemia [57]. In a series of 183 index patients from Erasmus MC, 21 COL4A1 and 3 COL4A2 variants were found mainly in children with porencephaly or other patterns of parenchymal haemorrhage, with 40% de novo variants (10/24) [58]. The 2020 consensus associates COL4A1 with porencephaly, schizencephaly, PMG and periventricular nodular heterotopia and recommends testing both genes together, although the evidence for COL4A2 is weaker [5]. The COL4A1 GeneReview (2016) does not mention schizencephaly or FCD, so this association rests on the primary series [27, 57].

2.8. Copy-number variants

The most frequent CNVs in MCD are 22q11.2 and 1p36 deletions (PMG), the 17p13.3 deletion with PAFAH1B1 and YWHAE (Miller-Dieker and isolated lissencephaly, 9% of a cohort of 811 patients with lissencephaly) and 6qter deletions (PMG, periventricular heterotopia) [5]. In Miller-Dieker syndrome, YWHAE is considered responsible for the facial features; the recurrence risk is negligible for de novo deletions and may be high when a parent carries a balanced rearrangement, and reciprocal 17p13.3 microduplications are associated with autism spectrum disorder, speech delay and ID, with distinct phenotypes depending on whether they include YWHAE (class I) or PAFAH1B1 (class II) [59]. The 22q11.2 deletion can unmask a recessive disorder on the non-deleted chromosome, such as CEDNIK syndrome (cerebral dysgenesis, neuropathy, ichthyosis and palmoplantar keratoderma), because SNAP29 lies within the typical deletion [55, 60]. The 1q43-q44 duplication with AKT3 causes macrocephaly with ID, and the same locus appears duplicated in mosaic form in HME and FCD [9, 41]. Several MCD genes have intragenic deletions and duplications detectable by microarray, and CNV mosaicism is detectable down to 15–20% [5].

2.9. Normalised reference variants

The descriptions were validated in VariantValidator (GRCh38, MANE Select) and ClinVar; the coding notations for PIK3R2 p.(Lys376Glu), PI4KA, ACTB and ACTG1 were obtained from ClinVar from the published protein notation [14, 61]. In MANE, the NEDD4L transcript is NM_001144967.3, not NM_015277.6 used in part of the literature [14]. Papers on GATOR1 use DEPDC5 NM_001242896, NPRL2 NM_006545 and NPRL3 NM_001077350 (GRCh37), and those on SLC35A2 use NM_005660, cited without version in the original sources; the SLC35A2 variants could not be validated in VariantValidator and keep the published notation, for example p.(Ser212Leufs*9), p.(Leu120Hisfs*7) and p.(Gln185*) [30, 35]. For somatic variants, the VAF does not equal the proportion of cells, except for X-linked genes in males [35].

2.10. Penetrance, expressivity and modifiers

With the exception of X-linked genes (ARX, DCX) and GATOR1 genes, MCD variants appear fully penetrant on imaging, but carriers may be asymptomatic and may never have had an MRI; parental MRI is therefore essential when a variant is inherited from an apparently unaffected parent, and female carriers of X-linked variants may have a normal MRI [5]. In GATOR1 epilepsies penetrance is reduced, estimated at about 60%, and expressivity varies in age at onset, seizure type, severity, response to treatment and presence of MCD [11, 30]. The GATOR1-specific variant classification framework uses a frequency threshold of 0.03% (83 alleles in gnomAD) for likely benign, calculated with a prevalence of genetic focal epilepsies of 0.32%, a DEPDC5 contribution of 9.4% and a penetrance of 60%, and a pathogenic allele count range of 6 (DEPDC5), 3 (NPRL2) and 4 (NPRL3) [30]. In the PI3K pathway megalencephalies, the predicted penetrance of germline variants is 100%, and phenotypic differences depend on the gene, the degree of activation, the timing of the post-zygotic variant and tissue distribution [9, 39, 45].

2.11. Class table: types, genes and discriminating tests

The table integrates the data described in this article and in the sources cited in each subsection [5, 7, 9–11, 13, 16].

3. Clinical Spectrum and Manifestations by System

3.1. Common presentation by age

MCD manifest differently depending on age. In the fetus, the signs are reduced fetal movements, polyhydramnios and anomalies on ultrasound or MRI; at birth, micro- or macrocephaly, dysmorphism and other congenital anomalies, abnormal tone and feeding and respiratory difficulties; in infancy, global developmental delay, hypo- or hypertonia, postnatal micro- or macrocephaly, cerebral palsy and epilepsy, including infantile spasms; in childhood, cerebral palsy, seizures, speech and cognitive delay, drooling or congenital suprabulbar paresis, visual defects and oculomotor apraxia; and in adolescence and adulthood, epilepsy, ID and tone abnormalities [5]. The presentation depends on extent and location: diffuse forms manifest early with severe delay, whereas small focal lesions may have focal epilepsy as their only manifestation, at any age [1, 6].

3.2. Polymicrogyria

PMG is one of the most frequent brain malformations, about 20% of all MCD, and is located in the Sylvian fissures in about 60–70% of cases, mainly in their posterior part, with frequent extension to the frontal, parietal and temporal lobes; the Sylvian fissures appear extended posteriorly and superiorly, with branching sulci [1, 13]. The topographic patterns described on axial T2-weighted MRI are as follows [13].

Bilateral perisylvian PMG is the most frequent pattern and the basis of congenital bilateral perisylvian syndrome: difficulties with tongue movements, expressive speech, sucking and swallowing, excessive drooling and facial diplegia, as in Worster-Drought syndrome, with mild to moderate ID in up to 75% and limb spasticity rarely severe [13]. In the multicentre study of 31 patients, all had facial, pharyngeal and masticatory diplegia (from mild dysarthria to inability to speak), ID (mild to severe) affected 85% and epilepsy 27 patients (87%), with atypical absences, atonic or tonic seizures and tonic-clonic seizures, poorly controlled in 55% [37]. PMG may be isolated or associated with grey matter heterotopia, ventriculomegaly and anomalies of the white matter, corpus callosum, brainstem and cerebellum [13]. Clinical severity depends on extent: focal forms may manifest only as epilepsy or learning difficulties, and extensive frontal PMG is often associated with spastic tetraparesis [8, 13]. In an individual patient data meta-analysis of 161 patients operated on for PMG-associated epilepsy, hemispheric PMG was associated with earlier seizure onset and a higher incidence of epileptic encephalopathy with spike-wave activation in sleep [62].

3.3. Schizencephaly

Schizencephaly has a broad anatomoclinical spectrum, with focal epilepsy in most patients and motor or cognitive deficit proportional to the extent and bilaterality of the clefts [16, 29]. The clefts are often multiple or accompanied by contralateral PMG, and anomalies of the fornix and septum pellucidum are common (“septo-optic dysplasia-plus”); open-lip schizencephaly may coexist with absence of the septum pellucidum and contralateral periventricular heterotopia [1]. Familial occurrence is rare [29]. In COL4A1-related disease, schizencephaly and porencephaly are accompanied by infantile hemiparesis, seizures, ID, dystonia, congenital cataract and, sometimes, haemolytic anaemia, raised CK and kidney injury [27, 57].

3.4. Focal cortical dysplasias

FCD may be located anywhere in the cortex, vary in size and be multilobar; type II is more frequent outside the temporal lobe, mainly in the frontal lobe [6]. Except in extensive FCD, there are no severe neurological deficits, and the main manifestation is epilepsy, with seizures at any age (usually in early childhood), very often drug-resistant and of high frequency, and with behavioural disturbances especially when epilepsy begins early [6]. FCD type II is the most frequent MCD in epilepsy surgery series (about 9% of all cases), with 51% in the frontal lobe and seizure onset at a mean of 5 years [7]. Fewer than 20% of patients with FCD respond to pharmacotherapy, and this response is transient [32]. The scalp electroencephalogram (EEG) shows focal rhythmic epileptiform discharges often correlated with the lesion, and stereo-EEG (SEEG) in FCD IIb shows absence of background activity, repetitive high-amplitude fast spikes followed by slow waves, interspersed with flat periods, and bursts of high-frequency oscillations that increase in slow-wave sleep and decrease in REM sleep [6]. Intrinsic epileptogenicity may not coincide with the MRI abnormality, and dysmorphic neurons, but not balloon cells, contribute to interictal spikes, fast gamma activity and ripples [7]. In a study of 38 MCD cases and 114 controls, repetitive epileptiform discharges and interictal focal fast discharges, and paroxysmal fast activity at seizure onset, were more prevalent in MCD [63].

In FCD type II, the OMIM clinical synopsis records severe, drug-resistant seizures, with multiple seizures per day (mainly focal impaired awareness seizures, often with secondary generalisation), focal neurological deficits such as hemiparesis, and cognitive impairment, with an MRI that may be normal [15]. In a surgical series of 34 cases (15 with balloon cells, IIB, and 19 with dysplasia only, IIA), IIB histology resembled that of TSC and IIA that of HME; almost one third of IIB cases had a normal MRI or very subtle changes, and patients with IIA had a more severe phenotype, with earlier onset, shorter seizure-free time after surgery, more hemiparesis and lower IQ [15]. In 21 operated adults, seizure onset occurred mainly in the third decade (18–55 years) and 16 (76%) became seizure-free after surgery, with equal outcomes in types IIA and IIB [15]. Type I has a less well-defined phenotype. In a series of 215 consecutive patients with FCD type I (Palmini classification), isolated forms (31%) had more frequent seizures, negative MRI, multilobar involvement and poorer postoperative control (46% Engel I), and forms associated with hippocampal sclerosis or tumours had the phenotype of the principal lesion [6]. FCD Ia was confirmed in a series of 19 children with early-onset seizures, subtle unilateral hemispheric hypoplasia of the posterior quadrant, global delay and drug resistance from onset (4% of 500 children operated on and 14.6% of all FCD), whereas FCD Ib and Ic still lack consistent clinicopathological characterisation [7]. FCD IIId with loss of layer 4 corresponds mainly to boys with occipital perinatal hypoxaemic injury, and Sturge-Weber syndrome is almost always accompanied by FCD IIIc [7]. mMCD is found in about 3% of operated patients, mainly adults, mostly with negative MRI [7]. FCD type I and mMCD are associated with prenatal and perinatal insults (severe prematurity, asphyxia, haemorrhage, hydrocephalus, stroke), and patients with these risk factors have more neurological abnormalities, lower intelligence quotient (IQ) and slower background EEG [4].

3.5. GATOR1 complex epilepsies

DEPDC5 variants cause a spectrum of epilepsy syndromes almost all with focal seizures: familial focal epilepsy with variable foci (FFEVF), autosomal dominant sleep-related hypermotor epilepsy (formerly autosomal dominant nocturnal frontal lobe epilepsy), familial mesial temporal lobe epilepsies, infantile spasms and, rarely, developmental encephalopathy with macrocephaly and PMG; in 2022 the ILAE defined the criteria for these syndromes of variable age at onset [11, 64]. About 60% of patients have a normal MRI, most have a family history of focal epilepsy and development is usually normal [11]. In FFEVF, each family member has a specific type of focal seizure, with a mean age at onset of 4.5 years; in sleep-related hypermotor epilepsy, nocturnal motor seizures are brief and stereotyped, occurring in clusters, with awareness often preserved, about 80% begin in the first two decades (mean 5 years) and the condition is lifelong, non-progressive and may attenuate in middle age [11].

The data are from the largest cohort described, referred by centres in 11 European and North American countries, with follow-up of 8 months to 58 years [30]. One third (22/73) had seizure onset in the first year of life, seizures predominated in sleep in 48%, the interictal EEG was epileptiform in 83%, only 11% became seizure-free with the first antiseizure medication, the mean number of drugs tried was four and cognitive or psychiatric comorbidity reached 60% (76% with onset in the first year) [30]. In the GeneReviews series (about 200 symptomatic heterozygotes), brain malformations occurred in 20% (FCD 17%, mainly type II; HME 2%; PMG in less than 1%), ID in 12% and ASD in 5%, ID was concentrated in patients with FCD or drug-resistant epilepsy, and sudden unexpected death in epilepsy (SUDEP) occurred in about 10% of families, even without cognitive impairment or polytherapy [11]. In the review of 183 probands, SUDEP occurred in 17 families (9.3%), whereas the overall incidence is 0.22/1000 per year in children and 1.2/1000 per year in adults with epilepsy; in a retrospective SUDEP cohort, DEPDC5 was found in about 10% (6/61) [30]. These figures document an association that justifies counselling and monitoring of risk factors, without quantifying an individual risk [16]. Biallelic missense variants of DEPDC5 (9 individuals from 5 families) cause a progressive developmental encephalopathy with early refractory multifocal epilepsy, bilateral PMG, congenital macrocephaly and ocular and cardiac defects [11].

3.6. Megalencephalies and hemimegalencephaly

MPPH

MPPH combines megalencephaly and bilateral perisylvian PMG. HC at birth ranges from normal to +6 SD and, in older individuals, from +3 to +10 SD; megalencephaly is congenital or early postnatal, rapidly progressive in the first year, and brain growth persists after ventricular shunting (true overgrowth) [9]. Fewer than 100 individuals have been described [9].

The GeneReview contains internally discordant percentages for epilepsy and oromotor dysfunction, and the table presents both values without choosing between them; early-onset epilepsy, focal or generalised, includes infantile spasms in some patients and may be refractory [9]. ID is determined by the extent and severity of the cortical malformation and by the age at onset and severity of epilepsy, being worse with neonatal and generalised onset [9]. The risk of hydrocephalus or of tonsillar ectopia with compression of the lower brainstem and upper spinal cord is greatest in the first 2 years, with limited retrospective data [9]. Visual problems (cortical visual impairment, blindness) are common, and fewer than 5 individuals each had congenital heart disease, hypoglycaemia, growth hormone (GH) deficiency, hypothyroidism, duplicated renal collecting system or medulloblastoma; three individuals with MPPH or a CCND2 variant developed medulloblastoma, suggesting a causal link [9]. CCND2 variants are associated with more extensive PMG (frontal or occipital), more severe epilepsy and ID and more frequent polydactyly, and AKT3 variants with connective tissue laxity and cutaneous capillary malformations overlapping with MCAP [9].

MCAP and the PROS spectrum

The criteria proposed in 2012 define MCAP as progressive megalencephaly with vascular anomalies or syndactyly, and MPPH as progressive megalencephaly with PMG, without the vascular anomalies or syndactyly of MCAP and without cerebral heterotopia [65]. MCAP includes macrocephaly, often with hydrocephalus that may require shunting, macrosomia, capillary malformation of the philtrum, reticulated capillary malformations (not true cutis marmorata), 2–3 toe syndactyly, postaxial polydactyly and asymmetry; perisylvian PMG is the classic brain phenotype, with ventriculomegaly, progressive tonsillar herniation due to rapid brain growth, Chiari I, a thickened corpus callosum and dilated dural venous sinuses [45, 66]. The variant is mosaic PIK3CA, not found in blood, and the condition belongs to the PIK3CA-related overgrowth spectrum (PROS) [16, 66].

Smith-Kingsmore syndrome

SKS combines variable overgrowth, developmental delay and ID, usually moderate to severe, autism, attention deficit hyperactivity disorder (ADHD) and congenital anomalies, without consensus criteria [10].

Independent walking is achieved at a mean of 2.5 years, but about one third do not achieve it, and about half use verbal communication; hyperpigmented naevi and pigmentary mosaicism are common in mosaic patients, and hypoglycaemia, present in some infants, usually resolves after the first year [10].

Hemimegalencephaly

HME is characterised by enlargement of one hemisphere, with abnormal cortical thickness and signal, white matter signal changes and an enlarged lateral ventricle; the septum pellucidum and fornices are often thickened or displaced, the ipsilateral cerebellum may be hypertrophied, and the lesion may extend to non-cerebral tissues [1, 5, 41]. Clinically, it combines ID and severe intractable epilepsy, which often requires removal or disconnection of the hemisphere, and it is associated with Proteus syndrome, other neurocutaneous syndromes and rarely with TSC [41]. The 2012 classification divides it into isolated, syndromic (neurocutaneous) and TSC-associated [4]. In a series of 15 infants operated on before 12 months (13 with HME), seizures began at a median of 3 days of life [67].

3.7. MOGHE and SLC35A2-associated epilepsy

MOGHE mainly affects young children with frontal lobe or “temporal plus” epilepsy, with a median onset at 2 years (0.3–13) [7]. In the combined series of 26 patients with MOGHE and a somatic SLC35A2 variant, onset occurred between 3 months and 7 years (mean 1 year and 3 months), with infantile spasms in 77% (20/26), often with tonic seizures, daily seizures in sleep and wakefulness in all, normal neurological examination in 85%, multiple EEG foci in 40%, a focal abnormality on presurgical MRI in 69%, focal hypometabolism on fluorodeoxyglucose positron emission tomography (FDG-PET) in 11/12 and presurgical cognitive impairment in 24/26 (mild to moderate in 63%, severe in 35%); the only difference from MOGHE without a variant was the higher frequency of spasms (77% versus 36%) [35]. In the cohort of 47 patients, the phenotype divided into early epileptic encephalopathy with spasms and moderate to severe ID (39 patients) and drug-resistant focal epilepsy with normal or borderline cognition and specific neuropsychological deficits (8 patients), with abnormal MRI in all of the former and half of the latter [51]. MRI varies with age: in young children, laminar T2/FLAIR hyperintensity at the corticomedullary junction (subtype I), and in older children reduced corticomedullary differentiation due to hyperintensity of the adjacent white matter (subtype II) [35]. Infantile epileptic spasms are the most common manifestation according to the 2026 review [52].

3.8. Heterotopias

Periventricular nodular heterotopia typically presents with focal epilepsy in women with normal cognition or learning difficulties, and the features suggesting FLNA are female sex, an X-linked dominant family history, absence of evident ID, bilateral confluent nodules along the frontal and central horns and a retrocerebellar cyst or mega cisterna magna; systemic manifestations include patent ductus arteriosus, aortic aneurysm, valvular dystrophy, obstructive lung disease, constipation, coagulopathy and hypermobility [5]. Posteriorly predominant periventricular heterotopia is often accompanied by overlying PMG, subcortical heterotopia and anomalies of the posterior fossa, corpus callosum and hippocampus [5]. In subcortical heterotopia of any cause (series of 107 individuals), 19% had normal development and 69% epilepsy, figures that do not apply to the EML1 form, in which 8 individuals from 5 families had megalencephaly, agenesis of the corpus callosum, severe delay, drug-resistant seizures, visual impairment and hydrocephalus in about half [16]. In NEDD4L-related periventricular nodular heterotopia, delay and ID are variable, refractory seizures occur in severe cases and there is marked intrafamilial variability [16].

3.9. Dysgyria, tubulinopathies and lissencephaly

In the tubulinopathies, motor and cognitive deficits are present in almost all patients and correlate with the severity of the malformation: lissencephaly, microlissencephaly and severe generalised dysgyria present with spastic tetraparesis and almost no voluntary motor control, whereas mild to moderate dysgyria presents with mild disability; epilepsy is very variable and is not determined by severity or by gene, and seizures occur in fewer than 30% of patients with a simplified gyral pattern [26]. Facial diplegia and pseudobulbar strabismus are frequent in centrally predominant pachygyria and in dysgyria [26]. Lissencephaly manifests in the newborn as poor feeding, hypotonia or seizures, with severe delay and early infantile spasms in most; severity and prognosis follow the Di Donato classification [8, 29]. In DYNC1H1-related disease, MCD (pachygyria, PMG or dysgyria in 48%, corpus callosum dysgenesis in 23%, heterotopia in 11%) accompany a motor neuropathy that may only be detectable electrophysiologically [54].

3.10. Other syndromes with MCD

Aicardi syndrome combines, mainly in girls, agenesis of the corpus callosum, chorioretinal lacunae and infantile spasms, with frontally and perisylvian-predominant PMG or pachygyria, periventricular and intracortical heterotopias, intracerebral cysts and choroid plexus papillomas; the combination of callosal agenesis, non-communicating cysts, subependymal heterotopia and PMG is relatively specific [68]. Baraitser-Winter cerebrofrontofacial syndrome (ACTB, ACTG1) combines dysmorphism, coloboma, deafness and frontally predominant pachygyria or SBH in 61% [69]. PI4KA-associated disease combines hypomyelinating leukodystrophy, cerebellar hypoplasia, intestinal atresias and immunodeficiency, and perisylvian PMG has only been described with the p.(Asp1854Asn) variant, including three fetuses with akinesia and arthrogryposis [70, 71]. Adams-Oliver syndrome (aplasia cutis and terminal limb defects; ARHGAP31, DOCK6, RBPJ, EOGT, NOTCH1, DLL4) may be associated with focal pachygyria, PMG, heterotopias and schizencephaly [72]. In Sturge-Weber syndrome, the leptomeningeal capillary malformation, caused by the somatic GNAQ variant p.(Arg183Gln) in 88% of the tissues assessed, is accompanied by cortical atrophy, gyriform calcification and occasional microgyria [73]. In linear sebaceous naevus syndrome, caused by somatic HRAS p.(Gly13Arg) or KRAS variants, CNS manifestations include micro- or macrocephaly, cerebral hypoplasia, arachnoid cysts, hydrocephalus, hemiparesis and calcifications [74].

4. Clinical Recognition and Red Flags

4.1. Prenatal period

Prenatal suspicion arises with reduced fetal movements, polyhydramnios, ventriculomegaly, micro- or macrocephaly, or gyration abnormalities on fetal ultrasound or MRI [5]. More than 60% of fetuses with dilated ventricles have another ultrasound anomaly, and the causes include disorders of neuronal migration and proliferation [75]. MCD due to under- or overmigration are difficult to recognise before 20 weeks, which is the legal limit for termination of pregnancy in many countries, and remain difficult afterwards: absence of normal sulci suggests the lissencephaly spectrum, early and aberrant sulcation suggests PMG, and heterotopias are often invisible [1]. Ventriculomegaly and microcephaly are detectable from 18 weeks with serial ultrasound scans, and lissencephaly only from 24 weeks and rarely prenatally [76]. PMG is detected on prenatal ultrasound in about 5% of cases, usually through microcephaly or associated malformations [13]. Ultrafast fetal MRI is used mainly to delineate intracranial anomalies, in particular migration defects, and in pregnancies at high risk of CNS malformation [76]. The EML1 form is detectable prenatally through ventriculomegaly and suspected migration disorder [16]. Any prenatal suspicion should be confirmed postnatally with ultrasound and MRI [1].

4.2. Newborn and infant

In the newborn, the signs that prompt investigation are micro- or macrocephaly, dysmorphism and malformations, abnormal tone, feeding and respiratory difficulties and neonatal seizures [5]. HME manifests with seizures in the first days of life, and MPPH with congenital or rapidly growing macrocephaly in the first year and hydrocephalus [9, 67]. In infants, infantile epileptic spasms should prompt a search for structural malformations (lissencephaly, heterotopia, FCD, HME, MOGHE) and TSC [35, 77]. Macrocephaly with capillary malformation of the philtrum, reticulated capillary malformations, syndactyly or asymmetry suggests MCAP; macrocephaly with postaxial polydactyly suggests MPPH; and macrocephaly with hyperpigmented macules or pigmentary mosaicism suggests SKS or mosaic MTOR-related MCD [9, 10, 66]. Severe congenital microcephaly with gyration abnormalities suggests microcephaly with lissencephaly or with PMG [1].

4.3. Child, adolescent and adult

In children, the red flags are oromotor dysfunction with drooling and dysarthria (bilateral perisylvian PMG), cerebral palsy without an explanatory perinatal history and epilepsy with developmental delay [5, 13, 78]. At any age, drug-resistant focal epilepsy requires a search for FCD with high-resolution MRI and an epilepsy protocol, especially when there are nocturnal hypermotor seizures or a family history of focal epilepsy suggestive of GATOR1 [11, 77]. In adulthood, mild forms (periventricular nodular heterotopia in women, focal PMG, small FCD) present as focal epilepsy or learning difficulties, and in familial PMG apparently minor findings (lisp, isolated learning difficulty) in close relatives justify MRI [5, 13].

4.4. Red flags and emergencies

The criteria for neurosurgery in MPPH, the risk of SUDEP in GATOR1 epilepsies, peroxisomal and infectious suspicion, precautions in COL4A1 disease and the visual toxicity of vigabatrin are documented in the sources cited [5, 9, 11, 27, 56, 79, 80].

4.5. Screening

There is no newborn screening for MCD. The dried blood spot card from newborn screening (Guthrie card) has retrospective diagnostic utility: testing for CMV DNA by PCR allows congenital infection to be distinguished from acquired infection when PMG or calcifications are diagnosed later [5, 56]. Children with pigmentary mosaicism, seizures and delay should be examined with a Wood's lamp for signs of TSC, incontinentia pigmenti or pigmentary mosaicism [81].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria and principles

There are no formal clinical criteria for most MCD: the definitive diagnosis is neuropathological, but tissue is rarely available (epilepsy surgery, autopsy), and in practice the diagnosis rests on imaging, complemented by clinical and genetic phenotyping [1, 13]. The consensus definitions in section 1.3 are the reference imaging criterion [5]. Criteria have been defined for MPPH (megalencephaly with PMG, with or without molecular confirmation; an undetected variant does not exclude the clinical diagnosis) and for MCAP (progressive megalencephaly with vascular anomalies or syndactyly), whereas SKS, GATOR1 epilepsies and MOGHE have no consensus criteria, and the diagnosis rests on the variant (SKS, GATOR1) or on histopathology (MOGHE) [7, 9–11, 65]. Imaging findings (generalised or focal, bilateral or unilateral) do not reliably distinguish genetic from non-genetic causes, and the consensus recommends seeking an aetiological diagnosis in all patients with MCD, because it explains the cause and ends the diagnostic odyssey, informs prognosis and recurrence risk, predicts the outcome of treatments such as epilepsy surgery and guides management (antivirals and hearing screening in congenital CMV, cardiovascular surveillance in FLNA- and ARFGEF2-related heterotopia, mTORC1 inhibition in TSC) [5].

5.2. Imaging protocol

MRI is the most important method; the best available equipment should be used, with 3 T preferable to 1.5 T when protocols are optimised, and the radiologist should receive from the clinician the HC, syndromic features, seizure semiology and the location of spikes on the EEG [1]. In the myelinated brain, the minimum protocol includes thin-section three-dimensional T1 (≤1 mm), axial and coronal T2 (≤3 mm) and FLAIR, preferably three-dimensional (≤1 mm) for triplanar reformatting; in the newborn, T2 in all three planes is essential and FLAIR is unnecessary; diffusion tensor imaging (DTI) is important for microscopic architecture and axonal orientation [1]. In epilepsy, MRI should follow an epilepsy protocol, and a negative MRI obtained at low field strength without a dedicated protocol is an inadequate acquisition; a large proportion of negative MRIs are due to substandard acquisitions and interpretation without the semiology and EEG [7, 77].

Myelination conditions interpretation: cortex–white matter contrast is maximal before myelination, decreases transiently during it and, in the T2 “isointense” phase from 8 to 12 months, may prevent recognition of an MCD; FCD may be visible in the newborn as T2-hypointense and T1-hyperintense, but from 2–3 months until the end of myelination may be almost indistinguishable [1]. Therefore, an MRI performed in the first year in patients with delay or seizures may need to be repeated after myelination is complete, in the third year of life; the Oxford Desk Reference indicates that an MRI performed before 15 months may need to be repeated, and the diagnostic consensus recommends repeating it when the first was performed between 3 months and 2.5 years or was of poor quality [1, 5, 81]. In PMG, high-quality MRI shows microgyri, microsulci and the “stippling” of the grey–white junction, specific to PMG, and the Sylvian fissures should be scrutinised in sagittal planes; before myelination the junction is thin (2–3 mm) and afterwards the cortex appears thicker and smoother (5–8 mm), never reaching the thickness of lissencephaly [1, 13]. MRI-measured cortical thickness should not be used as a rigid threshold, because it varies with age, myelination, sex, head size, area, lesions and seizures [1].

In FCD, the MRI signs are focal change in cortical thickness, abnormal gyral or sulcal pattern, cortical and subcortical signal changes, blurring of the grey–white junction and the “transmantle sign” (radial, funnel-shaped T2/FLAIR hyperintensity of the subcortical white matter extending to the ventricle), highly specific to FCD IIb, including tubers; asymmetric lack of subcortical myelination is fairly specific to FCD Ia, and focal enlargement of the subarachnoid spaces may point to the lesion [1, 6]. Almost one third of FCD type II remain MRI-negative even at 3 T, probably mainly type IIa; morphometric post-processing identifies subtle lesions, and 7 T MRI with post-processing detected subtle FCD in 22% of patients with focal epilepsy and negative 3 T MRI [7]. With strong clinical and electrical suspicion, apparently normal MRIs should be reviewed in multidisciplinary discussion [1].

Postnatal transfontanellar ultrasound assesses calcifications, gyration, ventriculomegaly and midline defects, but is operator-dependent; computed tomography (CT) is not recommended routinely for radiation protection reasons and remains second line for confirming calcifications, which susceptibility-weighted imaging (SWI) often detects [1]. In intracranial calcification, CT remains the reference standard when the pattern is unclear [56]. When an epileptogenic MCD is suspected with negative MRI, interictal FDG-PET co-registered with MRI (hypometabolism) is recommended and, if equivocal, ictal single-photon emission computed tomography (SPECT); peri-ictal arterial spin labelling perfusion MRI is promising [1, 7]. Proton spectroscopy is diagnostic only in a small number of metabolic disorders [46].

5.3. Diagnostic algorithm

The Neuro-MIG consensus proposes a flowchart applicable to any MCD; its steps, reconstructed from the text and the notes to the flowchart, are as follows [5]. First, expert review of the MRI, the prenatal and perinatal history (maternal rash or fever, twinning), a three-generation pedigree with consanguinity and clinical examination with HC, dysmorphism, skin and eyes. Second, consideration of an infectious cause when there are suggestive features (unwell infant, abnormal liver tests, retinal scars, deafness, calcifications, white matter injury or cysts). Third, broad genetic testing (CMA and NGS with analysis targeted to MCD genes; trio exome if negative), with genetic counselling, segregation and, if available, affected brain tissue, fibroblasts or saliva. Fourth, reassessment of the MRI to confirm that the variant explains the phenotype and, if negative, review of the classification and consideration of deep sequencing, another tissue or a non-genetic cause. The final clinical interpretation lies with the clinical geneticist, ideally after discussion with the molecular geneticist, neuroradiologist and neurologist [5]. Expert MRI review before genetic testing increased the yield to 37% in a mixed cohort of 117 patients, versus 18% in 784 patients without review (unpublished data from Regensburg), and almost doubled it in Rotterdam [5].

In PMG, HC is the first triage step [13]. Macrocephaly defines the group with the highest probability of a genetic cause, mainly of the PI3K–AKT–MTOR pathway: without systemic findings, MPPH (with shunted hydrocephalus or polydactyly in a subgroup); with somatic overgrowth or pigmentary changes, MCAP (vascular anomalies), MTOR (linear pigmentary changes) or PTEN (PHTS). Microcephaly with a fetal brain disruption pattern points to NDE1 or WDR62, and with complex additional malformations to pseudo-TORCH, RAB18 deficiency or tubulinopathies; normocephaly or borderline microcephaly point to CNVs (22q11.2, 1p36) [13]. In PMG, CMA remains the first genomic tier, and the yield of standard genomic tests is about 20%; suspicion of CMV leads to PCR on the Guthrie card, suspicion of a peroxisomal disorder to VLCFA, and megalencephaly to deep sequencing for mosaicism, as there are patients with mosaic PIK3R2 and normal HC [5].

When an FCD is found, the skin (hypomelanotic macules, shagreen patch) and the MRI (other FCD foci, subependymal nodules) are examined; if present, a full TSC work-up is performed, including TSC1 and TSC2 [5]. In families with focal epilepsy, GATOR1 single-nucleotide variants and CNVs are sought with autosomal dominant segregation with reduced penetrance; single-gene testing is rarely useful, and a panel (for example AKT3, CHRNA2, CHRNA4, CHRNB2, DEPDC5, KCNT1, MTOR, NPRL2, NPRL3, PIK3CA, PTEN, RHEB) or exome/genome is recommended, and, if there is no germline variant, methods for mosaicism and deep sequencing of tissue resected at surgery [5, 11]. When there is still no diagnosis, the consensus proposes visual inspection of the NGS reads, a search for deletions and duplications of the genes of interest, karyotype (86% consensus), single-nucleotide polymorphism microarray for regions of homozygosity in consanguineous families, metabolic investigation in microcephaly, PMG and cobblestone malformation, muscle biopsy if there is weakness or raised CK, repeat MRI, parental MRI, autopsy, DNA from brain tissue after surgery and trio genome with RNA sequencing within a research network [5].

5.4. Imaging pattern and targeted investigation

The table brings together the associations described in the consensus, the GeneReview and the Oxford Desk Reference [5, 7, 13, 29, 56]. The initial investigation of intracranial calcification includes congenital infection serology, CMV PCR on the Guthrie card, organic acids and amino acids, lactate, CSF (cells, protein, lactate, interferon alpha, pterins), genomic analysis and DNA storage [56]. Plasma CK and ophthalmological examination (retinal dysplasia, cataract, myopia) point towards the dystroglycanopathies, and nerve conduction studies are useful when there is hyporeflexia or foot deformities, suggestive of DYNC1H1, even when the question is a developmental one [16, 29]. Ophthalmological changes occur in up to 48% of patients with microcephaly, so ophthalmological examination is routine [5].

5.5. Test yield

The values are diagnostic yields in selected cohorts and not clinical sensitivities; they depend on the phenotype, MRI review and the tissue tested [5, 7, 53, 82]. In four lissencephaly patterns, the yield exceeds 90% for one or two genes (diffuse agyria with cortex thicker than 10 mm; occipital agyria with frontal pachygyria; temporal pachygyria with callosal agenesis and severe hypomyelination due to ARX; diffuse SBH with a band thicker than 5 mm), and a negative test in these patterns should lead to further investigation [5]. In the 2026 series, somatic variants explained 44% of FCDs, and germline NPRL3 (3) or DEPDC5 (2) variants were found in FCD IIa/IIb, with VAF detected down to 0.4% [82]. In DEPDC5 epilepsy, sequencing detects 96% of pathogenic variants and targeted deletion and duplication analysis 4%; in SKS, sequencing detects 100%, with no deletions or duplications described [10, 11]. In MPPH, 41 individuals with molecular confirmation were distributed among AKT3 (30%), CCND2 (30%) and PIK3R2 (40%) [9].

5.6. Mosaicism: tissue, depth and validation

Post-zygotic mosaicism has been described in the PI3K–AKT–mTOR pathway and in DCX, PAFAH1B1, FLNA and TUBB2B [5]. Standard blood exome sequencing often misses these variants; affected brain tissue is recommended and, in its absence, saliva or fibroblasts are preferable to lymphocytes, ideally with several tissues [5]. In SKS, 20–25% of variants are post-zygotic and it may be necessary to test more than one tissue, for example DNA from a fresh skin biopsy, with deep NGS, amplicons or droplet digital PCR (ddPCR) [10]. In the PIK3CA spectrum, DNA from affected tissue (fresh skin biopsy over an affected area, uncultured excised tissue) is tested with capture of the entire coding region and very deep NGS [16]. In MPPH, saliva or skin (affected or not) may detect a variant absent from blood, and CMA should include 1q43-q44 duplication/triplication [9].

The ILAE genetic protocol for surgical tissue (level 2 of the integrated diagnosis) specifies DNA from lesional tissue confirmed microscopically by the neuropathologist (frozen or formalin-fixed paraffin-embedded, FFPE), hybridisation capture and high-depth NGS (more than 1000×) of FCD genes, somatic variant callers and orthogonal validation by ddPCR or targeted amplicon sequencing; the nine genes of canonical FCD type II are AKT3, DEPDC5, MTOR, NPRL2, NPRL3, PIK3CA, RHEB, TSC1 and TSC2, and SLC35A2 should be included to distinguish MOGHE from FCD Ia [7]. The Neuro-MIG consensus adds that hybridisation is superior to amplicons, but that amplicons with unique molecular identifiers are effective, that the calling threshold should be adapted and that every low-level mosaic variant should be confirmed by an orthogonal technology [5]. Deep sequencing in unpaired FFPE tissue accurately detects low-level somatic variants, and two proof-of-concept studies detected brain variants in cell-free DNA from CSF, which could allow presurgical diagnosis if confirmed [5, 7]. Sampling determines detection: resected tissue corresponds to the epileptogenic zone, and depth, DNA quality and sample type (frozen or FFPE) alter sensitivity; in a patient with three operations, the SLC35A2 VAF was 2% in the tissue from the first and 22.6% in that from the second [16, 35].

5.7. Limitations of the tests

A negative result does not exclude a genetic cause. Panels and standard exome sequencing (mean coverage of 100 reads, minimum 30, calling threshold of 20% alternative reads) are adequate for germline variants and high-level mosaicism (more than 30% of cells), but not for low-level mosaicism [5]. Recommended coverage includes exons and 10 bp of intron (80% consensus), so deep intronic variants and regulatory variants, such as the 15 bp deletion of the ADGRG1 regulatory element in perisylvian PMG, escape standard testing [5, 16]. Commercial panels may not include all rare genes or detect mosaicism, and blood does not detect brain-restricted mosaicism or exclusively germline mosaicism [9, 13, 16]. A negative blood test does not exclude a brain-restricted mosaic MTOR variant in FCD or HME [16]. Balanced translocations and ring chromosomes require a karyotype, and homologous pseudogenes of some MCD genes require care with alignment [5]. A VUS neither establishes nor excludes the diagnosis [11].

5.8. Report content and interpretation

The report must state class 5 and 4 variants, all VUS (class 3) in MCD genes, and should consider rare de novo, homozygous or compound heterozygous variants in genes of uncertain significance; it should give the gene, transcript and version, HGVS nomenclature, zygosity, class according to the criteria of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) and whether exon-level CNV analysis was performed, recommend segregation and refer to a clinical geneticist; gene lists should be updated every 6 months and VUS shared in databases such as ClinVar and LOVD [5]. For somatic variants, the ILAE report should state the type of finding (gain or loss of function), the location, the sample (blood, FFPE or frozen) and the methodology, and distinguish a somatic variant (not inherited and not transmissible) from a germline one (possibly inherited and transmissible) [7].

The consensus proposes classifying as benign a variant with an allele frequency above 1% in gnomAD (instead of the ACMG 5%), considering exclusion of the gnomAD neurological cohort and taking homozygosity in five or more individuals as strong evidence of benignity, after excluding hemizygosity with a deletion [5]. The pathogenic variant TUBB2B p.(Ala248Val) appears in gnomAD with a frequency of 3% flagged as failed (artefact) and may appear in up to 30% of reads from controls in in-house data, which requires manual curation and Sanger sequencing of TUBB2B when the pattern is suggestive [5]. In the tubulinopathies, many variants are private and difficult to classify without functional studies, and the typical imaging pattern helps to interpret VUS [5]. In GATOR1 epilepsies, reclassification of 94 variants from the literature with the complex-specific framework moved 4 to likely benign, 12 to likely pathogenic and 17 to VUS, and variants inherited from unaffected parents may be causal, given incomplete penetrance [5, 30]. A germline GATOR1 variant does not exclude a somatic second hit, which explains the focal lesion and may require tissue to be demonstrated [11, 30].

5.9. Neuropathology and integrated diagnosis

Neuropathological examination, biobanking and genetic testing should accompany epilepsy surgery, autopsy and SUDEP, because there may be undiagnosed MCD; in 2016 the ILAE published a standard operating procedure for collecting and processing cortical samples, which should be anatomically intact, oriented and correlated with neurophysiology and MRI [5]. Histopathology is the gold standard for PMG, although clinical diagnosis rests on neuroimaging [13].

The scheme is that of the 2022 ILAE update [7]. In FCD type II, dysmorphic neurons are often cytomegalic (smallest diameter greater than 25 µm), glia are also dysmorphic, and the balloon cells of IIb, of mixed lineage, surround the area of dysmorphic neurons and do not generate abnormal electrical activity; non-phosphorylated neurofilament (SMI32) labels dysmorphic neurons in all FCD II, NeuN and MAP2 identify the architecture and heterotopic neurons, and OLIG2 the oligodendroglial hyperplasia of MOGHE [7]. Lentiform heterotopias of the superior temporal gyrus, classified as FCD IIIa in 2011, are a normal anatomical feature of the claustrum, and hypertrophic neurons should not be confused with dysmorphic neurons [7]. DNA methylation distinguishes epilepsy from controls and histopathological entities and may come to distinguish FCD subtypes [7].

5.10. Prenatal and preimplantation diagnosis

With the familial variant or variants known, prenatal diagnosis is performed by chorionic villus sampling at 11 weeks, and preimplantation genetic diagnosis is possible [9, 13, 29]. Ultrasound is not reliable for lissencephaly before 28 weeks, and fetal MRI is more accurate; associated features such as congenital microcephaly, hydrocephalus or cerebellar hypoplasia may give earlier clues [29]. In GATOR1 epilepsies, prenatal and preimplantation diagnosis is possible, but the phenotype cannot be predicted given reduced penetrance [11]. In forms due to brain-restricted somatic mosaicism (FCD, HME, MOGHE), the variant is not transmissible and prenatal diagnosis does not apply [7].

6. Differential Diagnosis

6.1. Comparative table

The table brings together the differentials described in the consensus sources, the GeneReviews and the Oxford Desk Reference [1, 5, 8, 12, 13, 16, 27, 56, 73].

6.2. Genetic or acquired cause

Congenital CMV is one of the most frequent non-genetic causes of MCD (PMG, calcifications, white matter changes, microcephaly) and accounts for 12–30% of PMG cases, or more when there are specific white matter changes [5]. In a cohort of 26 patients with bilateral PMG, 6 were CMV-positive (23%; the paper states 31%), without distinguishing prenatal from postnatal infection; in 50 patients with PMG, 6 (12%) were positive on the Guthrie card (unpublished data); and among 41 newborns with symptomatic CMV, 8 (19.5%) had microcephaly [5]. Infection should be suspected in the presence of microcephaly and congenital sensorineural deafness, white matter hyperintensity and calcifications; toxoplasmosis, syphilis, varicella-zoster and Zika produce similar pictures, and disruptive vascular events, maternal ergotamine and monozygotic twinning (death of a co-twin; twin–twin transfusion, with the donor twin more affected) are additional causes [5]. Dysmorphism, multiple anomalies, megalencephaly and microcephaly point to a genetic cause, although microcephaly also occurs in infection [5]. Calcifications suggest fetal brain injury with dystrophic calcification, but COL4A1/COL4A2 predisposes to fetal vascular injury and pseudo-TORCH syndrome mimics infection; in a series of 119 individuals with intracranial calcifications, there was a specific diagnosis in 50%, of whom 33 had Aicardi-Goutières, 6 OCLN-related pseudo-TORCH and 3 COL4A1 disease [5]. Congenital infections causing intracranial calcification also include rubella, parvovirus, herpes, lymphocytic choriomeningitis virus and HIV, and cystic leukoencephalopathy without megalencephaly due to RNASET2 may be indistinguishable from congenital CMV [56].

Pseudo-TORCH syndrome 1 due to biallelic OCLN manifests at or shortly after birth with seizures, feeding difficulties, severe progressive congenital microcephaly, tetraparesis and bulbar palsy, with continuous or semi-continuous ribbon-like cortical calcification, calcification of the thalamus, basal ganglia, pons and, in most of a large series, cerebellum, and bilateral symmetric frontoparietally predominant PMG; the calcification is mainly vascular, due to a blood–brain barrier defect, and the disease is enriched in consanguinity [16, 56].

6.3. PMG, pachygyria and cobblestone malformation

On low-resolution images, PMG and pachygyria may look similar, with an apparently thickened cortex and broad, smooth gyri; in PMG the cortex is overfolded but not thickened, and the microgyri, microsulci and stippling of the grey–white junction distinguish it [13]. The distinction between cobblestone malformation and PMG is difficult on low-resolution images and during myelination (3 months to 2 years): intracortical striations are regular and perpendicular in cobblestone malformation and chaotic in PMG, and hydrocephalus, brainstem hypoplasia and cerebellar cysts favour cobblestone malformation [5]. PMG on MRI may correspond to overmigration on histology, reflecting a common pathogenesis, and severe forms of cobblestone malformation have been incorrectly described as “pachygyria” [1, 5]. GPR56-related frontoparietal “PMG” and CHIME syndrome were described as PMG but are cobblestone malformations [5]. Severe congenital microcephaly with a reduced number of gyri, cobblestone malformation and PMG-like tubulin variants are the malformations most often confused with lissencephaly, with inefficient testing and incorrect counselling as a consequence; HC at birth above −3 SD with diffuse hypotonia favours lissencephaly, HC at or below −3 SD favours severe primary microcephaly, and spastic tetraparesis favours PMG, especially frontal [8]. The dysgyria of the tubulinopathies is distinguished by normal cortical thickness with sulci of abnormal depth and orientation and by dysmorphic basal ganglia [1, 26].

6.4. FCD, tubers, tumours and other focal lesions

The differential diagnosis of FCD includes low-grade tumours (gangliogliomas), “oligodendrocytosis” and MOGHE, whose abnormal signal lies in the seizure onset region [1]. TSC tubers are distinguished by the presence of subependymal nodules, SEGA and the other TSC criteria; the occasional calcification of FCD IIb appears at the base of tubers and is best seen on CT [1]. In temporal lobe epilepsy, temporopolar blurring and atrophy correspond to secondary white matter changes and not to FCD, and infiltrative glioneuronal tumours may mimic FCD IIIb [7]. FCD IIId should be distinguished from the occipital perinatal hypoxic–ischaemic lesions of which it is a consequence [7].

6.5. Periventricular heterotopia and subependymal nodules

TSC subependymal nodules are not isointense to grey matter, often calcify, may enhance with contrast when not calcified and are oriented perpendicular to the ventricular wall, whereas periventricular heterotopia is isointense to cortex on all sequences [1]. One or two isolated nodules with normal cognition have a low diagnostic yield, but may correspond to transmissible FLNA mosaicism [5].

6.6. Schizencephaly and porencephaly

Schizencephaly is a cleft lined by polymicrogyric grey matter, and porencephaly a clastic cavity lined by white matter and gliosis [1]. Porencephalic cysts may result from antenatal or neonatal parenchymal haemorrhagic infarction in neonatal alloimmune thrombocytopenia, coagulopathies (von Willebrand disease, factor V or X deficiencies), maternal warfarin exposure or thrombophilia, mainly heterozygous factor V Leiden; the genetic differentials of small vessel disease include COL4A2, CADASIL, TREX1-related retinal vasculopathy with leukoencephalopathy and CARASIL (HTRA1) [27]. Schizencephaly has genetic overlap with holoprosencephaly and porencephaly [55].

6.7. Inborn errors of metabolism with cortical malformation

The table transcribes the Saudubray classification [46]. In the PMG GeneReview, the metabolic causes listed are classic NKH, fumaric aciduria, glutaric acidaemia type II, histidinaemia, maple syrup urine disease, mitochondrial disorders, pyruvate dehydrogenase deficiency and Zellweger syndrome, Zellweger syndrome and fumaric aciduria being the most frequent, usually recognised by other features [13]. In D-bifunctional protein deficiency, PMG occurs in up to 65% of patients [5]. Fumarase deficiency may combine PMG and a thin or absent corpus callosum, peroxisomal disorders combine perisylvian PMG, pachygyria, hypomyelination and subependymal cysts, and glutathione peroxidase 4 (GPX4) deficiency causes severe pachygyria [46]. Congenital NAD deficiencies are a potential cause of major malformations that could be prevented with niacin, and the congenital microcephaly of serine synthesis defects is partially reversible with early treatment [46]. In disorders of cellular trafficking (cytoskeleton, golgipathies), the rule is an early-onset neurodevelopmental disorder with microcephaly, ID and white matter disease, and the tubulinopathies often combine cortical dysplasia and hypomyelination [83].

6.8. Holoprosencephaly (summary)

Holoprosencephaly (HPE) is the most frequent malformation of the prosencephalon and results from failure or incomplete division of the prosencephalon in the 3rd–4th weeks of gestation, affecting the hemispheres, deep structures and the olfactory and optic bulbs and tracts [12]. It is organised along a continuum of decreasing severity: alobar (monoventricle, no separation of the hemispheres), semilobar, lobar, middle interhemispheric fusion variant (syntelencephaly), septopreoptic type and microforms, which have craniofacial anomalies without a structural brain defect, such as a single central maxillary incisor [12]. It occurs in about 1/10,000 live births but in about 1/250 in early embryogenesis, because most fetuses are miscarried [84]. Between 25% and 50% of patients have a chromosomal anomaly, trisomy 13 being the most frequent cause (40–60% of all HPE), and CMA detects pathogenic CNVs in 10%; the non-syndromic form is autosomal dominant with highly variable expressivity, with SHH (5.4–5.9%), ZIC2 (4.8–5.2%), SIX3 (3%) and TGIF1 (less than 1%) [12]. MRI is the examination of choice, with review by a radiologist familiar with the subtypes [12].

6.9. Other structural malformations

Agenesis of the corpus callosum is one of the most frequent brain malformations and a less specific finding than is often thought, sometimes incidental with normal development; in a child with callosal agenesis, other midline anomalies, a more generalised migration defect (tubulins) or hydrocephalus (L1CAM) must be sought [55]. Septo-optic dysplasia combines optic nerve hypoplasia, agenesis of the septum pellucidum and hypothalamic–pituitary dysfunction (in one series, GH deficiency in 80%, ACTH deficiency in 60%, gonadotrophin deficiency in 40% and diabetes insipidus in 35%), is mostly sporadic and requires investigation of the entire hypothalamic–pituitary axis [55]. Arachnoid cysts are usually incidental, sporadic and single, more frequent in males, and bilateral temporal cysts should prompt testing for glutaric aciduria type 1 [55]. Cerebellar and pontocerebellar hypoplasias include microcephaly with pontocerebellar hypoplasia due to CASK and cerebellar hypoplasia due to VLDLR [85].

7. Treatments and Risk Management Strategies

7.1. Principles and sources of recommendations

There is no comprehensive management guideline for MCD: management is rarely evidence-based, the Neuro-MIG consensus statements deal with diagnosis and classification, and the GeneReviews for MPPH, SKS and DEPDC5 epilepsy state that there are no formal guidelines and that the surveillance recommendations are provisional [5, 9–11]. No drug has an approved indication for isolated FCD, HME, PMG, MOGHE or GATOR1 epilepsy, and the use of mTOR inhibitors in these conditions is off-label, extrapolated from the indication approved by the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) of the European Union (EU) for seizures in TSC, which shares the pathway, and from the trials and series described below [16, 31]. The cross-cutting sourced recommendations are the ILAE definition of drug resistance, the ILAE recommendations for seizures in infancy and for paediatric presurgical evaluation, and the genetic recommendations of the ACMG and the American Academy of Pediatrics (AAP) [79, 86–89]. Treatment is organised by a paediatric neurologist or neurologist experienced in multiple disability and refractory epilepsy, with an individual plan from diagnosis [26].

7.2. Assessment after diagnosis

The table summarises the initial assessments proposed by the GeneReviews for the entities of the class [9–11, 27, 29].

7.3. Epilepsy: principles of pharmacological treatment

Drug-resistant epilepsy is defined as failure of two tolerated, appropriately chosen and used antiseizure medication schedules, as monotherapy or in combination, to achieve sustained seizure freedom, and is the criterion for referral for surgical assessment [86]. In MCD, antiseizure medications are used according to seizure type and standardised protocols; many may be effective and none has been demonstrated specifically in these disorders, and rapid control matters, because poorly controlled seizures worsen feeding and increase the need for gastrostomy and the risk of aspiration [9, 11, 26]. In GATOR1 epilepsy, only 11% became seizure-free with the first antiseizure medication, the mean number of drugs tried was four and 68% received three or more; of those who became seizure-free on monotherapy, 5 of 8 were taking sodium channel blockers, and the ketogenic diet, used in 5 patients, improved 2 [30]. In MPPH, one patient with AKT3 and refractory spasms responded to the ketogenic diet, an anecdotal observation that should not be presented as an expected response [9, 16]. For seizures in infancy, the ILAE recommends MRI as standard investigation at tertiary level (level A), referral to tertiary level after failure of one antiseizure medication (ideally after the first seizure) and identification of surgical candidates as part of standard practice at primary and secondary levels, with surgical screening in experienced tertiary centres; levetiracetam has strong evidence in focal seizures of infancy [79]. The dosing of broad-spectrum antiseizure medications follows the Summary of Product Characteristics (SmPC) of each drug; the doses below refer to drugs with a specific relationship to MCD or the mTOR pathway.

7.4. Infantile epileptic spasms

Infantile epileptic spasms occur in HME, MOGHE, extensive FCD, GATOR1 epilepsies (6–10%) and diffuse MCD, and infantile epileptic spasms syndrome has been defined by the ILAE since 2022 [30, 35, 90]. The ILAE (2015) considers adrenocorticotrophic hormone (ACTH) preferable for short-term control (level B), oral corticosteroids probably effective (level C) and a shorter interval between onset and treatment possibly beneficial for neurodevelopment (level C) [79]. The 2012 American Academy of Neurology guideline, now retired, considered low-dose ACTH probably as effective as high-dose and more effective than vigabatrin in the short term outside TSC, and serves only as historical context; in TSC, vigabatrin is first line [91].

The doses of vigabatrin and corticotropin are those of the US labels and the European SmPC [79, 80, 92, 93]. The depot corticotropin preparation of the US label has no EMA-authorised equivalent in the sources consulted.

7.5. Epilepsy surgery

Resective surgery is the treatment with the highest probability of seizure freedom in drug-resistant focal MCD, and should be explored early in FCD and HME [7, 11]. The 2014 ILAE consensus recommendations on tests in the presurgical evaluation of children with drug-resistant focal epilepsy, based on expert opinion owing to the lack of class 1–2 evidence, consider surgery very effective in selected children [87]. Outcomes depend on histological type and genetic cause: excellent in FCD II, unfavourable in FCD I, and in FCD III the same as those of the principal lesion; in FCD IIb and bottom-of-sulcus FCD, localisation and resection can be performed with excellent results without extraoperative invasive EEG, with resection of the gyrus down to the bottom of the sulcus guided by intraoperative electrocorticography with depth electrodes [7]. MCD due to mTOR pathway genes (germline or germline with a second hit) have better surgical outcomes than MCD due to ion channel and synaptic transmission genes, and a germline GATOR1 variant is not a contraindication to surgery [7, 30].

The table brings together the results described in the sources, which are retrospective series without a comparison group, with different follow-up periods and outcome definitions [6, 7, 11, 30, 35, 51, 62, 67, 73, 82, 94]. Complete and extensive resection improves the outcome in MOGHE: in the series of 20 patients, all with extensive resections achieved Engel I, and in the Cleveland series success was associated with complete resection of the sampled ictal onset regions and near-total lobectomy [7, 94]. In the SLC35A2 cohort, the probability of Engel I decreased with epilepsy duration and increased with a normal or improved postoperative EEG, and cognition remained unchanged in most after surgery [51]. In PMG, Engel I was associated with shorter epilepsy duration (8.05 versus 11.92 years, p = 0.009), and SEEG was valuable for functional mapping and tailored resections [62]. In FCD type I, even invasive techniques often fail to localise the epileptogenic zone, owing to more diffuse pathology [7]. In bilateral perisylvian PMG, callosotomy improved seizures in several patients in the 1993 study [37].

In HME, hemispherectomy or hemispheric disconnection is often necessary; the unaffected hemisphere is smaller than that of controls, which may explain postoperative seizure control and a cognitive outcome below expectations [4, 41]. In the Warsaw series (2005–2024), the 15 infants were operated on at a median of 6.4 months (minimum weight 6400 g); all required intraoperative transfusion (median 44.8 mL/kg; one third more than one blood volume), with no mortality, with intracerebral haemorrhage in 13.3%, intraoperative cardiac arrest in 6.7%, intensive care stay longer than 5 days in 40% and shunting for hydrocephalus in 26.7% [67]. About one third of patients with FCD continue to have seizures after resection with a complete work-up [32].

7.6. Therapy targeting the mTOR pathway

Everolimus

Everolimus (Votubia®, EU; Afinitor Disperz®, USA) is an mTORC1 inhibitor approved as adjunctive therapy for focal seizures associated with TSC from 2 years [31, 95]. In the USA, the recommended starting dose is 5 mg/m² orally once daily, titrated to trough concentrations of 5–15 ng/mL, with new dose = current dose × (target concentration/current concentration) and a maximum increment of 5 mg per adjustment [31]. In the EU, the starting dose depends on age and inducers: under 6 years, 6 mg/m² (9 mg/m² with a CYP3A4/P-glycoprotein inducer); from 6 years, 5 mg/m² (8 mg/m² with an inducer), titrated to the same trough range in increments of 1–4 mg [95]. The trough is measured 1–2 weeks after starting, after a dose or formulation change, and 2 weeks after starting or stopping moderate CYP3A/P-glycoprotein inhibitors, oral cannabidiol or strong inducers, or after a change in hepatic function; thereafter, every 3–6 months if body surface area changes and every 6–12 months if it is stable [31, 95]. Grade 2 non-infectious pneumonitis requires interruption until grade 0–1 and resumption at 50%, and grade 2 stomatitis interruption until grade 0–1 and resumption at the same dose, with reduction to 50% if it recurs [31]. In the pivotal EXIST-3 trial, stomatitis affected 55% and 64% of patients in the low- and high-trough arms (versus 9% with placebo), hypercholesterolaemia 86% and 85% (versus 58%) and neutropenia 25% and 37% (versus 23%), with discontinuation due to adverse effects in 5% and 3% [31].

In MCD not associated with TSC, everolimus is off-label. In a series of 5 patients with drug-resistant GATOR1 epilepsy (median 18 seizures per month; 5–16 previous failed antiseizure medications) treated with everolimus titrated to 5–15 ng/mL, the 3 with DEPDC5 loss of function had seizure reductions of 74.3–86.1% (one stopped at 12 months because of psychiatric symptoms), the patient with a DEPDC5 missense variant had 43.9% and the patient with an NPRL3 splicing variant worsened; stomatitis was the most frequent adverse effect [96]. In another series of 4 patients with NPRL2 or NPRL3 and intractable focal seizures (15–301 seizures per month; 6–14 failed antiseizure medications), with daily doses of 3.75–11.5 mg and troughs of 5–8.7 ng/mL, 2 became seizure-free at 2 months (one with relapse when the trough fell below 4 ng/mL and renewed freedom with higher levels) and 2 had reductions of 52% and 86%; there was stomatitis in 2 (discontinuation in 1), hyperlipidaemia in 1 and recurrent respiratory infections in 1, resolved with dose reduction [97]. Response appears to depend on the gene and the variant, and worsening has been described [16]. A randomised crossover trial from Yonsei University (NCT03198949), with everolimus 4.5 mg/m²/day titrated to a trough of 5–15 ng/mL in 23 patients aged 4–40 years with FCD type II refractory after surgery, ended in 2020 with no results in the registry and no publication identified in the searches performed [98].

Sirolimus

Sirolimus (Rapamune®) is approved in the USA for prophylaxis of renal transplant rejection (from 13 years) and for lymphangioleiomyomatosis (LAM), and in the EU for prophylaxis of renal rejection in adults and for sporadic LAM with moderate lung disease or declining function; it has no indication in MCD [99, 100]. In LAM, the starting dose is 2 mg/day, with the trough measured at 10–20 days and adjustment to 5–15 ng/mL, maintaining each dose for 7–14 days before a further adjustment because of the long half-life; the label sets no maximum dose in LAM, and in transplantation the dose given on any one day should not exceed 40 mg; it is given once daily, consistently with or without food, and the tablets should not be crushed or split; the label carries a boxed warning of immunosuppression, infection and lymphoma, and safety under 13 years has not been established in transplantation [99]. In the FCDS-01 trial in FCD type II, the starting dose was 1 mg/day below 40 kg and 2 mg/day from 40 kg, once daily, with adjustment to a trough of 5–15 ng/mL over 8–24 weeks (plus 1 mg/day every 2 weeks until week 8 and then every 4 weeks) and 12-week maintenance with 2 or 4 mg/day, with no maximum dose defined in the source; mean maintenance troughs were 4.6–5.0 ng/mL, below target [32]. All patients had adverse effects, mainly stomatitis and infections; the pneumonitis marker KL-6 rose (median 170.5 to 243 U/mL), with 2 patients above 500 U/mL without pneumonitis and 1 grade 2 pneumonitis, and there were adjustments for anorexia, somnolence and proteinuria; testosterone was monitored in boys over 9 years because of the risk of male infertility [32]. In SKS, sirolimus was used off-label in 7 individuals at doses of 0.5–2 mg/day, with parent-reported improvements in verbal communication, attention, self-injury and hyperphagia; troughs above 3 ng/mL were associated with sleep–wake cycle changes, and the GeneReview suggests considering a lower dose with a target below 3 ng/mL and monitoring sleep, without endorsing the treatment [10]. The SKS dose (target below 3 ng/mL) and that of the FCD trial (5–15 ng/mL) differ because they seek different effects and tolerate different toxicities.

The table brings together the dosing in the sources cited, and off-label use in MCD results from extrapolation of the TSC indication [10, 31, 32, 95–99]. Before using an mTOR inhibitor, the SKS GeneReview recommends assessing the biochemical and functional impact of the variant and points out that first-, second- and third-generation inhibitors have distinct mechanisms and off-target effects; the FCDS-01 trial suggests genotyping the tissue before treatment, because the expected response is greater with an mTOR pathway variant [10, 32]. The University of Washington Rx-Matcher platform collects off-label treatment experiences [10].

7.7. Other antiseizure medications with a related indication

Cannabidiol (Epidyolex®, EU; Epidiolex®, USA) is approved for TSC seizures. In the USA, from 1 year, the starting dose is 2.5 mg/kg twice daily (5 mg/kg/day), with weekly increases of 2.5 mg/kg twice daily up to the recommended maintenance of 12.5 mg/kg twice daily (25 mg/kg/day), which the label gives as the maximum recommended human dose in TSC [101]. In the EU, from 2 years, treatment starts with 2.5 mg/kg twice daily in the first week, moves to 5 mg/kg twice daily (10 mg/kg/day) and the response is assessed, with weekly increases up to a maximum of 12.5 mg/kg twice daily (25 mg/kg/day), weighing benefit and risk above 10 mg/kg/day [102]. Transaminases and bilirubin are measured before starting; in the TSC trial, discontinuation due to adverse effects was 11% with 25 mg/kg/day versus 3% with placebo, and the effects in at least 10% were diarrhoea, raised transaminases, decreased appetite, somnolence, pyrexia and vomiting [101]. Starting or stopping cannabidiol requires measuring the everolimus trough 2 weeks later [31]. Extension to patients with MCD outside TSC has not been studied.

Cenobamate (Ontozry®, EU; Xcopri®, USA) is approved for focal seizures in adults, in the USA as monotherapy or add-on and in the EU as add-on after at least two antiseizure medications; titration is 12.5 mg/day in weeks 1–2, 25 mg in weeks 3–4, 50 mg in weeks 5–6, 100 mg in weeks 7–8, 150 mg in weeks 9–10 and 200 mg/day maintenance from week 11, with a maximum of 400 mg/day in increments of 50 mg every 2 weeks; titration should not be accelerated because of the risk of drug reaction with eosinophilia and systemic symptoms (DRESS), the maximum is 200 mg/day in mild to moderate hepatic impairment and the drug shortens the QT interval [103, 104]. Its use in adults with FCD or GATOR1 epilepsy is an extrapolation from the focal seizure indication.

7.8. D-galactose in MOGHE and SLC35A2-CDG

Oral D-galactose seeks to bypass the UDP-galactose transport defect and is not authorised as a medicine for this indication. In the open-label pilot trial NCT04833322, 12 patients with histological MOGHE (5–28 years, three centres), with seizures or cognitive impairment and epileptiform activity after surgery, received oral D-galactose for 6 months at doses up to 1.5 g/kg/day; 6 had a brain somatic SLC35A2 variant [33]. Tolerability was good (abdominal discomfort in 2, resolved by spacing or reducing the dose), 9 of 12 were global responders (6 of 6 with SLC35A2), with a seizure reduction of at least 50% in 3 of 6 patients assessed, EEG improvement in 2 of 5 and 1 patient seizure-free, and improvements in impulsivity, social communication and executive function; the authors conclude that it is safe and call for larger studies [33]. In germline SLC35A2-CDG, an open-label trial with doses up to 1.5 g/kg/day for 18 weeks in 10 patients (NCT02955264) showed improvement in the Nijmegen Paediatric CDG Rating Scale (NPCRS) (p = 0.008) and in glycosylation [16]. One patient treated for 7 years improved from an NPCRS score of 22 to 12 with normalisation of glycans, but 7 of 8 additional patients in the same publication had no evident clinical or biochemical response [105]. A randomised crossover trial with medical-grade D-galactose (AVTX-801) in SLC35A2-CDG is planned (NCT05402384) [106].

7.9. Hydrocephalus, tonsillar ectopia and other complications

In MPPH, neurosurgical assessment is required when HC increases rapidly, there is obstructive hydrocephalus, signs of raised intracranial pressure or progressive or symptomatic tonsillar ectopia/Chiari; early treatment of hydrocephalus may reduce the risk of progressive ectopia, but data on the best management are lacking [9]. Cerebellar overgrowth in the first 2 years explains acquired tonsillar ectopia, and in MCAP hydrocephalus may require shunting [1, 45]. Heart disease, thyroid and renal changes and medulloblastoma are treated by the respective specialties, and the hypoglycaemia of SKS usually resolves after the first year [9, 10]. In SKS, obesity is treated with lifestyle and nutritional intervention (one person successfully treated with semaglutide), and sleep with sleep medicine, behavioural therapy and possibly pharmacotherapy [10].

7.10. Feeding, rehabilitation and development

The oromotor dysfunction of perisylvian PMG and MPPH requires dietary modification, speech therapy and, sometimes, gastrostomy; in severe malformations a nasogastric tube and then gastrostomy are used [9, 26]. Physiotherapy prevents contractures, scoliosis and hip dislocation, and early prevention of contractures reduces the risk of pressure ulcers and difficulties with mobility and hygiene; occupational therapy, vision therapy for oculomotor deficits and strabismus, and augmentative and alternative communication, which does not impair speech development, complete rehabilitation [9, 26]. Early intervention from 0 to 3 years and adapted special education follow diagnosis, and ASD and ADHD receive standard treatment, including applied behaviour analysis [9, 11]. Polydactyly is corrected surgically, and the congenital external ophthalmoplegia of the tubulinopathies may require surgical or non-surgical treatment [9, 26].

7.11. Regulatory status

US dates come from Drugs@FDA and the labels, and EU dates from EMA lists and procedural steps taken after authorisation documents [31, 80, 92, 93, 95, 100–104, 108–112]. The correspondence between the 2018 efficacy supplement for Afinitor Disperz and the TSC seizure indication is inferred from the date and the label, because the Drugs@FDA summary does not describe the indication [108]. National availability and reimbursement in Portugal and Brazil are not stated in the sources consulted.

7.12. Clinical trials and studies

The sources are the labels, papers and registrations cited [16, 31–33, 96–98, 113, 114]. The trials cannot be compared with each other: they differ in population (TSC versus isolated FCD versus GATOR1), design (randomised with placebo versus single-arm or series), duration and definition of response [31, 32]. In FCDS-01, estimated seizure frequencies fell from 32.3 (90% CI 14.8–70.7) at baseline to 17.4 in weeks 5–8 (p = 0.003), but the primary outcome did not reach significance; the prospective external control (RES-FCD, n = 60, median reduction of 0.5%) was not comparable (propensity scores 0.02 versus 0.67), troughs were below target, and the authors suggest that higher doses could be beneficial, by analogy with EXIST-3, in which the high trough had more response [32]. The relative reduction in EXIST-3 should be read together with the absolute response (40% versus 15.1% responders) and the stomatitis rate (64% versus 9%) [31]. The GATOR1 series included 9 patients, without controls, and showed outcomes ranging from reductions of more than 80% to worsening [96, 97]. The randomised everolimus trial in FCD type II has not published results, and a search without results does not prove that none exist [98].

7.13. Developments in the last 6–12 months

The sources are the registrations, the sponsor press release and the publications cited [48, 52, 53, 62, 67, 82, 94, 105, 106, 115–119]. None of these studies has yet produced an authorisation or regulatory opinion; the announcement of the first patient dosed with paxalisib is a sponsor press release, and a 2026 preprint on clinical deep sequencing of mosaics in MCD has not been peer-reviewed [116, 120]. The 2026 ERN EpiCARE position paper identifies as priorities prevention of epileptogenesis and disease modification, genetics and targeted therapies, surgical decision-making, trial designs and outcome measures, and comorbidities and mortality [34].

7.14. Agents and situations to avoid

In COL4A1 disease, smoking and hypertension, which increase the risk of stroke, prolonged head compression during delivery, activities with a risk of head trauma, and anticoagulants should be avoided [27]. Vigabatrin is not recommended in the EU in the presence of a pre-existing clinically significant visual field defect [92]. mTOR inhibitors interact with inhibitors and inducers of CYP3A and P-glycoprotein and with cannabidiol, which requires measuring the trough after each change [31]. In DEPDC5 epilepsy, everolimus was associated with worsening in a patient with an NPRL3 variant, and response to the inhibitor should be assessed individually [96].

7.15. Pregnancy, lactation and the perioperative period

Women with epilepsy have higher mortality in pregnancy, and antiseizure medications reduce this risk; fetal exposure may increase risk depending on the drug, dose and stage of gestation, but is often lower than that of untreated maternal seizures, so the decision should be discussed before conception, with transition to a lower-risk drug when possible and counselling on increasing periconceptional folic acid to 4000 µg per day and on the effect of pregnancy on the metabolism of antiseizure medications [11]. When the fetus is at risk of COL4A1 disease, caesarean section is recommended to prevent cerebrovascular injury due to birth trauma [27]. In hemispheric surgery in infants with HME, blood loss requires preparation for transfusion (median 44.8 mL/kg in the Warsaw series) and prolonged intensive care in 40% [67]. Sirolimus has a boxed warning of immunosuppression and infection risk, to be considered in the perioperative period [99].

mTOR inhibitors require specific contraceptive and reproductive planning, described in the labels and SmPCs and summarised in the following table.

The recommendations are those of sections 8.1–8.3 of the US labels and section 4.6 of the European SmPCs [31, 95, 99, 121]. The breastfeeding recommendation for sirolimus differs: the US label calls for individual weighing, and the European SmPC requires breastfeeding to be discontinued [99, 121]. The labels do not include pregnancy data in women with MCD; monitoring of testosterone in boys over 9 years in the sirolimus trial in FCD reflects the same risk to male fertility [32].

7.16. Transition and palliative care

In SKS, transition to adult care is prepared from 10 years of age [10]. In severe malformations, the level of care in serious intercurrent illness should be discussed with the family, and care coordination, palliative care and home nursing form part of the plan when needed [9, 26].

7.17. Guideline recommendations and discrepancies

The recommendations reproduce the sources cited [5, 7, 11, 32, 79, 86–89, 91]. The vigabatrin labels differ in the visual monitoring intervals, and those for everolimus and cannabidiol in the starting and maintenance doses between the USA and the EU [31, 80, 92, 95, 101, 102].

8. Reproductive Options and Family Counselling

8.1. Principles

Counselling in MCD depends on three decisions: the reliability of MRI interpretation, because PMG is easily confused with cobblestone malformation and empirical estimates require correct classification; the distinction between a germline variant (possibly inherited and transmissible) and a somatic variant restricted to one tissue (neither inherited nor transmissible); and the exclusion of non-genetic causes, such as congenital infection, and of chromosomal anomalies [5, 7, 29]. When there is no molecular diagnosis, the patient's DNA should be stored [9, 56].

8.2. Empirical recurrence risks by entity

The table reproduces the empirical risks of the Neuro-MIG consensus, which require highly reliable MRI interpretation [5]. In PMG without a family history or causal variant, the risk is uncertain; the Oxford Desk Reference cites an apparently higher risk in perisylvian PMG, especially in boys, of up to 30%, and points out that many cases are sporadic [29]. In schizencephaly, familial occurrence is rare [29]. In an isolated girl with learning difficulties and macrocephaly without a diagnosis, the risk for siblings is about 3–5%, with attention to dominant disorders with variable expressivity [45].

8.3. Counselling by entity

In MPPH, inheritance is autosomal dominant, almost always due to a de novo germline variant; there is somatic mosaicism for PIK3R2 and AKT3 in some patients, vertical transmission of PIK3R2 from an affected parent to several children in one family, and parental germline mosaicism suggested in three families [9]. Parents should be assessed with targeted testing and neurological examination with HC measurement, because a parent with somatic and germline mosaicism may be mildly affected and leucocyte testing may detect neither somatic mosaicism nor exclusively germline mosaicism; the risk for siblings is 50% if a parent is a carrier and slightly higher than that of the population if the variant is not detected, and the risk for the children of a patient is 50% (germline) or lower (mosaic) [9]. In the PIK3CA spectrum, vertical transmission and recurrence in siblings have not been confirmed; with a mosaic proband, the risk for siblings is that of the population, the theoretical risk due to parental germline mosaicism when the variant is constitutional is below 1%, and the risk for children is below 50% [16].

In SKS, about 86% of variants are de novo constitutional or early post-zygotic, about 11% are inherited from an unaffected parent with gonadal mosaicism and about 3% from an affected parent; recurrence in siblings due to presumed gonadal mosaicism has been described in several families, so the risk for siblings is higher than that of the population even when the variant is not detected in the parents, whereas a proband with a post-zygotic variant allows the parents to be presumed non-carriers [10].

In GATOR1 epilepsies, many patients inherit the variant from an often asymptomatic parent, with penetrance that may be as low as 60% and variable expressivity; the risk for siblings is 50% if a parent is a carrier and the risk for children 50%, with no prediction of phenotype [11]. Some patients with MCD have a germline variant and a mosaic second hit in the brain only, and parental gonosomal mosaicism has been described [11]. The rare biallelic form follows the recessive risk of 25%, 50% and 25%, and there is one report of epilepsy risk in heterozygous parents [11]. The GeneReview recommends clarifying the genetic status of older and younger asymptomatic relatives to identify early those at risk of seizures, and accepts predictive testing in asymptomatic adults after the variant has been identified [11]. In asymptomatic minors, targeted testing is justified by the seizure surveillance it allows; the sources consulted do not address the minimum age for predictive testing.

In the tubulinopathies, more than 95% of variants are de novo; rarely a parent is affected (mainly TUBB3), and maternal germline mosaicism has been described in two families with several affected children, so the risk for siblings is slightly higher than that of the population when the parents are negative [26]. In KIF5C, the variant p.(Glu237Val) in 4 affected siblings with presumed maternal germline mosaicism means that the recurrence risk cannot be said to be negligible [16]. In recessive PMG (ADGRG1, OCLN, WDR62, NDE1, dystroglycanopathies), the risk for siblings is 25%, and in X-linked PMG the mother of more than one affected son is an obligate heterozygote; with a negative mother in a simplex case, the risk for siblings is slightly higher than that of the population (below 1%) owing to germline mosaicism [13]. In COL4A1 disease, inheritance is autosomal dominant with 40% de novo variants in one series, and reduced penetrance and phenotypic variation make counselling difficult; the first manifestations may appear in asymptomatic adults and an abnormal MRI may be silent [27, 58]. In FLNA-related periventricular nodular heterotopia and the DCX spectrum, counselling follows X-linked inheritance [29].

In somatic forms restricted to the brain (FCD type II due to MTOR, PIK3CA, AKT3 or RHEB; HME; MOGHE due to SLC35A2), the variant is not transmissible and the recurrence risk is that of the population, provided that a germline variant in a repressor of the pathway has been excluded [5, 7].

8.4. Parental testing and cascade testing

When an MCD variant is inherited from an apparently unaffected parent, parental MRI is essential, because, except for X-linked and GATOR1 genes, carriers always show structural changes, even when asymptomatic [5]. In familial PMG without a genetic cause, MRI of close relatives with neurological problems or apparently minor findings may reveal the familial form [13]. Segregation should be recommended in the report when the parents or siblings have not been tested [5].

8.5. Prenatal and preimplantation diagnosis

Once the familial variant or variants have been identified, prenatal diagnosis (chorionic villus sampling at 11 weeks) and preimplantation genetic diagnosis are possible, and, when the variant is not known, fetal imaging has the limitations described in section 5.10 [9, 13, 29]. In cases of possible germline mosaicism (MPPH, SKS, tubulinopathies, KIF5C), prenatal diagnosis should be discussed even when the variant was not found in the parents [9, 10, 16, 26]. In GATOR1 epilepsies, the phenotype of a carrier fetus cannot be predicted [11].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Frequency and what each figure measures

The table distinguishes proportions in clinical or surgical series, birth prevalences and counts of published cases, which are not interchangeable [5, 7, 8, 11, 13, 16, 18, 20–23, 28, 68, 84]. In the European Epilepsy Brain Bank, seizure onset was before 18 years in 75.9%, but 72.5% were operated on as adults, with a mean epilepsy duration before surgery of 20.1 years in adults and 5.3 years in children; FCD was the most frequent diagnosis in children and hippocampal sclerosis in adults [28]. ClinVar counts (for example, 340 pathogenic, 164 likely pathogenic and 1223 VUS variants in DEPDC5) reflect panel volume and not clinical frequency [16].

9.2. United States and Europe

US population data include the birth prevalence of schizencephaly (1.54/100,000), the population-based California study (about 4 million births between 1984 and 2001) that associated schizencephaly with young maternal age, monozygotic twinning and vascular disruption, and the incidence estimates for Aicardi syndrome [4, 18, 68]. European data come mainly from surgical series and consortia: the European Epilepsy Brain Bank, the GATOR1 cohort referred by centres in Belgium, Denmark, France, Germany, Italy, Switzerland, Czechia, the Netherlands and the United Kingdom (as well as Canada and the USA), the MOGHE cohorts from Germany, Paris, Amsterdam and Seoul, and the Neuro-MIG network laboratories, with 737 patients of in-house yield [5, 28, 30, 35]. In Europe, the birth prevalence of schizencephaly has been validated only for the United Kingdom [18].

9.3. Founder populations and other regions

The deletion of exons 3 and 4 of OCLN is a founder allele in the Egyptian population, and pseudo-TORCH syndrome is enriched in consanguinity [16]. Fukuyama congenital muscular dystrophy, with cobblestone malformation, is comparatively common in Japan owing to a founder effect (3 kb insertion in FKTN), and muscle–eye–brain disease is comparatively common in Finland [29]. In the recessive forms (ADGRG1, WDR62, ASPM), consanguinity increases the diagnostic probability: in consanguineous families, ASPM and WDR62 explain more than 50% of primary microcephalies, and 14 of the 30 patients with bifrontoparietal PMG in one series came from 8 consanguineous families [5, 16]. The PI4KA series includes two Turkish individuals homozygous for p.(Asp1854Asn) [70]. COL4A1 disease has been described in families of Dutch, Italian, French, German, American, Chinese, Spanish and Japanese origin [16]. An Argentine cohort of 38 patients with lissencephaly, SBH or periventricular heterotopia had a yield of 36% with targeted testing [5].

9.4. Portugal and Brazil

The sources consulted do not include epidemiological data on MCD from Portugal or Brazil; in 2018 Lisbon hosted consensus meetings of the Neuro-MIG network that produced the imaging and diagnostic recommendations used in this article [1, 5]. The absence of data in the sources consulted does not demonstrate their absence from the literature.

9.5. Special populations

In children, FCD is the most frequent lesion in epilepsy surgery, and the diagnosis depends on myelination maturity, which requires repeating the MRI after the third year when the first was negative [1, 28]. In infants undergoing hemispheric surgery before 12 months, all required transfusion and 40% required intensive care for more than 5 days [67]. In adults, MRI-negative FCD type II and mMCD predominate in surgical series, and the duration of epilepsy before surgery is long [7, 28]. In women with epilepsy of childbearing age, pregnancy management follows section 7.15 [11]. Patients with severe ID need adapted visual assessment on vigabatrin, because perimetry is only reliable with a developmental age above 9 years [92].

9.6. European reference networks and patient associations

ERN EpiCARE (https://epi-care.eu/) is the European reference network for rare and complex epilepsies, and ERN ITHACA (https://ern-ithaca.eu/) brings together the centres for rare malformation syndromes and ID [34]. The Neuro-MIG research network (https://www.neuro-mig.org/) produced the European consensus statements on classification and diagnosis [1, 5]. Patient associations and resources listed in the sources include the Polymicrogyria Awareness Organization (www.pmgawareness.org), FOCUS Families for septo-optic dysplasia (www.focusfamilies.org), the Epilepsy Foundation, the American Epilepsy Society, the National Institute of Neurological Disorders and Stroke, the American Association on Intellectual and Developmental Disabilities and MedlinePlus [11, 13, 26, 55].

10. Illustrative Clinical Case

10.1. Hemimegalencephaly due to a brain-restricted somatic AKT3 variant

Poduri and colleagues described in 2012 a non-dysmorphic boy with seizures from the first week of life, caused by right HME [41]. Preoperative MRI, on coronal and axial T2-weighted sections, showed enlargement of the right hemisphere, abnormal cortical thickness and signal, abnormal white matter signal and an enlarged lateral ventricle (Figure 1G–H of the paper), and the patient underwent hemispherectomy at 5 months; postoperative MRI showed CSF in place of the resected hemisphere (Figure 1I–J) [41]. Sequencing of DNA from the resected brain tissue identified the AKT3 variant NM_005465.7:c.49G>A p.(Glu17Lys) in 8 of 46 clones (17.4%), corresponding to heterozygosity in about 35% of cells, and the variant was not detectable in leucocyte DNA (Figure 3D, electropherograms from brain and leucocytes) [14, 41]. After surgery, he had two periods of breakthrough seizures; at 10 years he had been seizure-free for 6 years, with left hemiparesis, independent walking and fluent speech, and he read and attended school with special education support [41]. The figures consulted are available at https://pmc.ncbi.nlm.nih.gov/articles/PMC3460551/figure/F1/ (pre- and postoperative MRI) and https://pmc.ncbi.nlm.nih.gov/articles/PMC3460551/figure/F3/ (molecular analysis) [41].

The case illustrates three general points of the class. The variant was present only in the affected tissue, so germline testing in blood would have been negative and the molecular diagnosis was only possible with the surgical tissue [5, 41]. The somatic variant is not transmissible, which allows the parents to be reassured about recurrence risk [7]. And early hemispheric surgery was followed by prolonged seizure freedom and motor autonomy, a better outcome than the average described in series of hemispheric surgery in infants (Engel I in 46% of HMEs at 12 months in the Warsaw series); the case is an individual report and does not represent the expected course [67]. There is no clinical evidence for pathway inhibition targeting AKT3, and no inhibitor is approved in HME [16].

10.2. Complementary case

In the same paper, another non-dysmorphic boy (HMG-1) with left HME underwent hemispherectomy at 15 months; seizures fell from about 10 per day to 1–4 per month, and at 6 years he had right hemiparesis, independent walking, good comprehension and speech limited to a few words, attending school with support [41]. The brain tissue showed mosaic trisomy of chromosome 1q in about 40% of cells (copy number 2.41), absent from blood, with the histology in Figure 2 (https://pmc.ncbi.nlm.nih.gov/articles/PMC3460551/figure/F2/) [41]. This case shows that a mosaic CNV can cause the same phenotype as a point variant, which justifies copy-number analysis in resected tissue [5, 41].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis by entity

The prognosis of MCD depends on the extent of the malformation, the age at onset and severity of epilepsy, and the cause [1, 9]. In bilateral perisylvian PMG, ID is mild to moderate in up to 75% and motor dysfunction is rarely severe, but epilepsy occurs in about 87% and is poorly controlled in more than half in a classic series [13, 37]. In MPPH, ID is determined by the extent of the malformation and by the age at onset and severity of epilepsy, and the risk of complications from hydrocephalus and tonsillar ectopia is concentrated in the first 2 years [9]. In SKS, survival into adulthood is common (one individual alive at 70 years, with a variant in the FIT domain), about one third do not achieve independent walking and about half use verbal communication, with some early deaths [10]. In the tubulinopathies, mild forms survive into adulthood and severe forms may lead to death at a young age from seizures or pneumonia [26]. In lissencephaly, life expectancy relates to severity on neuroimaging, with 50% mortality at 10 years in grade 3 with a normal cerebellum and much higher with cerebellar hypoplasia [8].

In GATOR1 epilepsies, cognition is normal in 54% of probands, drug resistance reaches 54% and SUDEP occurred in 9.3% of the families of 183 probands; sleep-related hypermotor epilepsy is lifelong, but may attenuate in middle age [11, 30]. In operated FCD type II, 67.4% are free of disabling seizures at 5 years and 39.4% are off antiseizure medications; in mMCD, 45% are seizure-free at 2 years; and in isolated FCD type I postoperative control is poorer [6, 7]. In MOGHE with SLC35A2, surgery led to Engel I–II in 77% in one series and Engel I in 63.8% in another, with cognition preserved in most; longer epilepsy duration was associated with a poorer outcome [35, 51]. In HME, hemispheric surgery before 12 months achieved Engel I in 46% at 12 months [67]. Outcome data come from retrospective surgical series and not from untreated natural history, which is not described in the sources consulted for FCD and MOGHE.

11.2. Surveillance

The GeneReviews surveillance recommendations are provisional and are not derived from formal guidelines; the drug monitoring intervals are those of the labels and match section 7 [1, 5, 9–11, 31, 80, 92, 95, 99, 101]. In the trials of mTOR inhibitors in FCD, monitoring included KL-6 as a marker of pneumonitis and testosterone in boys over 9 years [32]. In COL4A1 disease, the initial assessment includes MRI, CT angiography, ophthalmology and renal and hepatic imaging, and follow-up is guided by the clinical course, avoiding vascular risk factors [27]. Epilepsy follow-up is by neurology, and patients with drug-resistant FCD or HME should be referred early to an epilepsy surgery centre [11, 79].

12. Summary and Final Considerations

MCD are a heterogeneous group of disorders of cerebral cortex formation, classified by the stage of development primarily affected (proliferation, migration, post-migrational development) and, increasingly, by the biological pathway involved. The imaging definitions of the Neuro-MIG consensus, the 2012 Barkovich classification and the ILAE classification of FCD updated in 2022, with mMCD, MOGHE and the three-layer integrated diagnosis, form the common basis for diagnosis.

MRI with an age-adapted protocol, reviewed by an expert, is the first examination, and incomplete myelination requires repeating the MRI after the third year when the first was negative or early. HC and skin examination guide investigation: macrocephaly points to the PI3K–AKT–mTOR pathway, microcephaly to proliferation genes, calcifications to congenital infection or genetic mimics, and clefts to vascular disruption and collagen IV. CMA and broad NGS are the baseline genetic tests, but a negative result in blood does not exclude a genetic cause: FCD type II, HME and MOGHE often result from brain-restricted somatic variants, which only deep sequencing of the affected tissue detects, and germline GATOR1 and TSC variants are accompanied by a somatic second hit in the lesion.

Treatment remains mainly symptomatic. Epilepsy surgery offers the highest probability of seizure freedom in FCD type II, MOGHE, HME and some PMG, with outcomes depending on histological type, genetic cause, extent of resection and duration of epilepsy; a germline GATOR1 variant does not contraindicate surgery. mTOR inhibitors have an approved indication for TSC seizures, and their use in isolated FCD, GATOR1 epilepsies and SKS is off-label, supported by a single-arm trial that did not meet its primary outcome and by small series with heterogeneous responses; D-galactose in MOGHE and SLC35A2-CDG, paxalisib and radiprodil are under investigation.

Genetic counselling depends on the distinction between germline variants, mosaics with a risk of transmission and non-transmissible brain somatic variants, on the reduced penetrance of GATOR1 and on the germline mosaicism described in MPPH, SKS and the tubulinopathies, and requires reliable MRI interpretation. The ERN EpiCARE and ERN ITHACA networks and the Neuro-MIG research network structure access to centres with experience in the diagnosis and surgery of these malformations.

13. References

1. Severino M, Geraldo AF, Utz N, Tortora D, Pogledic I, Klonowski W, et al. Definitions and classification of malformations of cortical development: practical guidelines. Brain. 2020;143(10):2874-2894. doi:10.1093/brain/awaa174.
2. Orphanet. Genetic cerebral malformation. ORPHA:269553. https://www.orpha.net/en/disease/detail/269553
3. Orphanet. Cerebral cortical dysplasia. ORPHA:268950. https://www.orpha.net/en/disease/detail/268950
4. Barkovich AJ, Guerrini R, Kuzniecky RI, Jackson GD, Dobyns WB. A developmental and genetic classification for malformations of cortical development: update 2012. Brain. 2012;135(Pt 5):1348-1369. doi:10.1093/brain/aws019.
5. Oegema R, Barakat TS, Wilke M, Stouffs K, Amrom D, Aronica E, et al. International consensus recommendations on the diagnostic work-up for malformations of cortical development. Nat Rev Neurol. 2020;16(11):618-635. doi:10.1038/s41582-020-0395-6.
6. Blümcke I, Thom M, Aronica E, Armstrong DD, Vinters HV, Palmini A, et al. The clinicopathologic spectrum of focal cortical dysplasias: a consensus classification proposed by an ad hoc Task Force of the ILAE Diagnostic Methods Commission. Epilepsia. 2011;52(1):158-174. doi:10.1111/j.1528-1167.2010.02777.x.
7. Najm I, Lal D, Alonso Vanegas M, Cendes F, Lopes-Cendes I, Palmini A, et al. The ILAE consensus classification of focal cortical dysplasia: An update proposed by an ad hoc task force of the ILAE diagnostic methods commission. Epilepsia. 2022;63(8):1899-1919. doi:10.1111/epi.17301.
8. Brock S, Dobyns WB, Jansen A. PAFAH1B1-Related Lissencephaly / Subcortical Band Heterotopia. 2009 Mar 3 [updated 2021 Mar 25]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK5189/
9. Mirzaa G. MPPH Syndrome. 2016 Nov 17 [updated 2022 Jul 28]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK396098/
10. Prada CE, Raski CR, Mirzaa G, Besterman A, Krueger DA, Liu AC. Smith-Kingsmore Syndrome. 2025 Nov 20. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK619247/
11. Baulac S, Baldassari S. DEPDC5-Related Epilepsy. 2016 Sep 29 [updated 2023 Mar 9]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK385626/
12. Tekendo-Ngongang C, Muenke M, Kruszka P. Holoprosencephaly Overview. 2000 Dec 27 [updated 2020 Mar 5; revised 2026 Jun 9]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1530/
13. Stutterd CA, Dobyns WB, Jansen A, Mirzaa G, Leventer RJ. Polymicrogyria Overview. 2005 Apr 18 [updated 2018 Aug 16; retired 2024 Jun 6]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1329/
14. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-68. doi:10.1002/humu.23348.
15. Online Mendelian Inheritance in Man (OMIM). #607341 Focal cortical dysplasia, type II; FCORD2 (entry and clinical synopsis). Baltimore (MD): Johns Hopkins University; created 12-11-2002, updated 06-04-2017. Available from: https://www.omim.org/entry/607341 [accessed 26 October 2026].
16. GeneGuidance. Clinical curation of the genes MTOR, DEPDC5, NPRL2, NPRL3, PIK3CA, AKT3, PIK3R2, CCND2, ADGRG1, TUBB2B, EMX2, COL4A1, SLC35A2, NEDD4L, EML1, OCLN, KIF5C, DYNC1H1, WDR62, DCX, PAFAH1B1, TSC1 and TSC2 [local database]. Version of 26 September 2026 [accessed 26 October 2026].
17. Orphanet. Bilateral frontoparietal polymicrogyria. ORPHA:101070. https://www.orpha.net/en/disease/detail/101070
18. Orphanet. Schizencephaly. ORPHA:799. https://www.orpha.net/en/disease/detail/799
19. Orphanet. Polymicrogyria. ORPHA:35981. https://www.orpha.net/en/disease/detail/35981
20. Orphanet. Isolated focal cortical dysplasia. ORPHA:65683. https://www.orpha.net/en/disease/detail/65683
21. Orphanet. Hemimegalencephaly. ORPHA:99802. https://www.orpha.net/en/disease/detail/99802
22. Orphanet. Epilepsy-mild cortical malformation-oligodendroglial hyperplasia. ORPHA:729224. https://www.orpha.net/en/disease/detail/729224
23. Orphanet. Megalencephaly-polymicrogyria-postaxial polydactyly-hydrocephalus syndrome. ORPHA:83473. https://www.orpha.net/en/disease/detail/83473
24. Orphanet. Periventricular nodular heterotopia. ORPHA:98892. https://www.orpha.net/en/disease/detail/98892
25. Orphanet. Pseudo-TORCH syndrome type 1. ORPHA:1229. https://www.orpha.net/en/disease/detail/1229
26. Bahi-Buisson N, Maillard C. Tubulinopathies Overview. 2016 Mar 24 [updated 2021 Sep 16]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK350554/
27. Plaisier E, Ronco P. COL4A1-Related Disorders. 2009 Jun 25 [updated 2016 Jul 7]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK7046/
28. Blumcke I, Spreafico R, Haaker G, Coras R, Kobow K, Bien CG, et al. Histopathological Findings in Brain Tissue Obtained during Epilepsy Surgery. N Engl J Med. 2017;377(17):1648-1656. doi:10.1056/nejmoa1703784.
29. Lissencephaly, polymicrogyria, and neuronal migration disorders. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 216–221.
30. Baldassari S, Picard F, Verbeek NE, van Kempen M, Brilstra EH, Lesca G, et al. The landscape of epilepsy-related GATOR1 variants. Genet Med. 2019;21(2):398-408. doi:10.1038/s41436-018-0060-2.
31. Novartis Pharmaceuticals Corporation. AFINITOR (everolimus) tablets; AFINITOR DISPERZ (everolimus tablets for oral suspension). Prescribing information. NDA 022334/203985. Revised 2026 Jun [effective 2026-06-01]. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=2150f73a-179b-4afc-b8ce-67c85cc72f04
32. Kato M, Kada A, Shiraishi H, Tohyama J, Nakagawa E, Takahashi Y, et al. Sirolimus for epileptic seizures associated with focal cortical dysplasia type II. Ann Clin Transl Neurol. 2022;9(2):181-192. doi:10.1002/acn3.51505.
33. Aledo-Serrano Á, Valls-Carbó A, Fenger CD, Groeppel G, Hartlieb T, Pascual I, et al. D-galactose Supplementation for the Treatment of Mild Malformation of Cortical Development with Oligodendroglial Hyperplasia in Epilepsy (MOGHE): A Pilot Trial of Precision Medicine After Epilepsy Surgery. Neurotherapeutics. 2023;20(5):1294-1304. doi:10.1007/s13311-023-01395-z.
34. Tchaicha S, Auvin S, Beniczky S, Brunklaus A, Lagae L, Perucca E, et al. Areas of research priorities in epilepsy: A position paper of the European Reference Network for Rare and Complex Epilepsies, EpiCARE. Epilepsia Open. 2026;11(1):291-321. doi:10.1002/epi4.70210.
35. Bonduelle T, Hartlieb T, Baldassari S, Sim NS, Kim SH, Kang HC, et al. Frequent SLC35A2 brain mosaicism in mild malformation of cortical development with oligodendroglial hyperplasia in epilepsy (MOGHE). Acta Neuropathol Commun. 2021;9(1):3. doi:10.1186/s40478-020-01085-3.
36. Taylor DC, Falconer MA, Bruton CJ, Corsellis JA. Focal dysplasia of the cerebral cortex in epilepsy. J Neurol Neurosurg Psychiatry. 1971;34(4):369-387. doi:10.1136/jnnp.34.4.369.
37. Kuzniecky R, Andermann F, Guerrini R. Congenital bilateral perisylvian syndrome: study of 31 patients. The CBPS Multicenter Collaborative Study. Lancet. 1993;341(8845):608-612. doi:10.1016/0140-6736(93)90363-l.
38. Piao X, Hill RS, Bodell A, Chang BS, Basel-Vanagaite L, Straussberg R, et al. G protein-coupled receptor-dependent development of human frontal cortex. Science. 2004;303(5666):2033-2036. doi:10.1126/science.1092780.
39. Rivière JB, Mirzaa GM, O'Roak BJ, Beddaoui M, Alcantara D, Conway RL, et al. De novo germline and postzygotic mutations in AKT3, PIK3R2 and PIK3CA cause a spectrum of related megalencephaly syndromes. Nat Genet. 2012;44(8):934-940. doi:10.1038/ng.2331.
40. Lee JH, Huynh M, Silhavy JL, Kim S, Dixon-Salazar T, Heiberg A, et al. De novo somatic mutations in components of the PI3K-AKT3-mTOR pathway cause hemimegalencephaly. Nat Genet. 2012;44(8):941-945. doi:10.1038/ng.2329.
41. Poduri A, Evrony GD, Cai X, Elhosary PC, Beroukhim R, Lehtinen MK, et al. Somatic activation of AKT3 causes hemispheric developmental brain malformations. Neuron. 2012;74(1):41-48. doi:10.1016/j.neuron.2012.03.010.
42. Dibbens LM, de Vries B, Donatello S, Heron SE, Hodgson BL, Chintawar S, et al. Mutations in DEPDC5 cause familial focal epilepsy with variable foci. Nat Genet. 2013;45(5):546-551. doi:10.1038/ng.2599.
43. Ishida S, Picard F, Rudolf G, Noé E, Achaz G, Thomas P, et al. Mutations of DEPDC5 cause autosomal dominant focal epilepsies. Nat Genet. 2013;45(5):552-555. doi:10.1038/ng.2601.
44. Lim JS, Kim WI, Kang HC, Kim SH, Park AH, Park EK, et al. Brain somatic mutations in MTOR cause focal cortical dysplasia type II leading to intractable epilepsy. Nat Med. 2015;21(4):395-400. doi:10.1038/nm.3824.
45. Macrocephaly. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 224–227.
46. Saudubray JM, García-Cazorla Á. Clinical approach to inborn errors of metabolism in paediatrics. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1. doi:10.1007/978-3-662-63123-2_1.
47. Ribierre T, Deleuze C, Bacq A, Baldassari S, Marsan E, Chipaux M, et al. Second-hit mosaic mutation in mTORC1 repressor DEPDC5 causes focal cortical dysplasia-associated epilepsy. J Clin Invest. 2018;128(6):2452-2458. doi:10.1172/jci99384.
48. Maletic M, Bizzotto S, Ribierre T, Guerdoud K, Raoux C, Doladilhe M, et al. Mosaic human cortical organoids model mTOR-related focal cortical dysplasia through DEPDC5 deletion. Brain. 2026;149(10):3631-3646. doi:10.1093/brain/awag086.
49. D'Gama AM, Woodworth MB, Hossain AA, Bizzotto S, Hatem NE, LaCoursiere CM, et al. Somatic Mutations Activating the mTOR Pathway in Dorsal Telencephalic Progenitors Cause a Continuum of Cortical Dysplasias. Cell Rep. 2017;21(13):3754-3766. doi:10.1016/j.celrep.2017.11.106.
50. Mirzaa GM, Conti V, Timms AE, Smyser CD, Ahmed S, Carter M, et al. Characterisation of mutations of the phosphoinositide-3-kinase regulatory subunit, PIK3R2, in perisylvian polymicrogyria: a next-generation sequencing study. Lancet Neurol. 2015;14(12):1182-1195. doi:10.1016/s1474-4422(15)00278-1.
51. Barba C, Blumcke I, Winawer MR, Hartlieb T, Kang HC, Grisotto L, et al. Clinical Features, Neuropathology, and Surgical Outcome in Patients With Refractory Epilepsy and Brain Somatic Variants in the SLC35A2 Gene. Neurology. 2023;100(5):e528-e542. doi:10.1212/wnl.0000000000201471.
52. Spyrou J, Lockhart PJ, Maljevic S, Howell KB, Reid CA. Mild malformation of cortical development with oligodendroglial hyperplasia in epilepsy (MOGHE): genetics, mechanisms and precision therapy. Acta Neuropathol. 2026;152(1):14. doi:10.1007/s00401-026-03059-6.
53. Lai A, Neil JE, Akula SK, Amrom D, Andermann E, Bergin A, et al. Diverse Genetic Etiologies of Unilateral Polymicrogyria. Ann Neurol. 2026;99(5):1277-1286. doi:10.1002/ana.78169.
54. Möller B, Coppola A, Jungbluth H, Salimi Dafsari H. DYNC1H1-Related Disorders. 2024 Mar 21. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK601997/
55. Structural intracranial anomalies (agenesis of the corpus callosum, septo-optic dysplasia, and arachnoid cysts). In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 320–323.
56. Intracranial calcification. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 200–203.
57. Yoneda Y, Haginoya K, Kato M, Osaka H, Yokochi K, Arai H, et al. Phenotypic spectrum of COL4A1 mutations: porencephaly to schizencephaly. Ann Neurol. 2013;73(1):48-57. doi:10.1002/ana.23736.
58. Meuwissen ME, Halley DJ, Smit LS, Lequin MH, Cobben JM, de Coo R, et al. The expanding phenotype of COL4A1 and COL4A2 mutations: clinical data on 13 newly identified families and a review of the literature. Genet Med. 2015;17(11):843-853. doi:10.1038/gim.2014.210.
59. Jones KL, Jones MC, del Campo M. Miller-Dieker syndrome (lissencephaly syndrome). In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 258–261.
60. 22q11 deletion syndrome. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 624–627.
61. National Center for Biotechnology Information. ClinVar [Internet]. Bethesda (MD): NCBI. Available from: https://www.ncbi.nlm.nih.gov/clinvar/ (VCV000039814, VCV000039816, VCV000039808, VCV000280374, VCV000217823, VCV000160146, VCV000864862, VCV000208451, VCV000208450, VCV000127168, VCV000803470)
62. Bosisio L, Viglietti C, Consales A, Nobile G, Pacetti M, Tortora D, et al. Surgical treatment of epilepsy in polymicrogyria: A subject-level meta-analysis and decision-making framework. Epilepsia. 2026;67(7):3304-3317. doi:10.1002/epi.70204.
63. Shakhatreh L, Quiles LEP, Chen Z, Gonen OM, Mohidat S, Millist L, et al. Interictal and seizure-onset scalp electroencephalographic patterns in malformations of cortical development. Epilepsia. 2026;67(6):2979-2991. doi:10.1002/epi.70193.
64. Riney K, Bogacz A, Somerville E, Hirsch E, Nabbout R, Scheffer IE, et al. International League Against Epilepsy classification and definition of epilepsy syndromes with onset at a variable age: position statement by the ILAE Task Force on Nosology and Definitions. Epilepsia. 2022;63(6):1443-1474. doi:10.1111/epi.17240.
65. Mirzaa GM, Conway RL, Gripp KW, Lerman-Sagie T, Siegel DH, deVries LS, et al. Megalencephaly-capillary malformation (MCAP) and megalencephaly-polydactyly-polymicrogyria-hydrocephalus (MPPH) syndromes: two closely related disorders of brain overgrowth and abnormal brain and body morphogenesis. Am J Med Genet A. 2012;158A(2):269-291. doi:10.1002/ajmg.a.34402.
66. Jones KL, Jones MC, del Campo M. Megalencephaly-capillary malformation-polydactyly syndrome. In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 732–735.
67. Pielas M, Kotulska K, Roszkowski M, Syczewska M, Jurkiewicz E, Kowalczyk P. Hemispheric surgery for hemimegalencephaly and hemispheric cortical dysplasia in infants below 12 months of age-A single-center experience. Epilepsia. 2026;67(8):3996-4007. doi:10.1002/epi.70294.
68. Sutton VR, Van den Veyver IB. Aicardi Syndrome. 2006 Jun 30 [updated 2020 Nov 12]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1381/
69. Verloes A, Drunat S, Pilz D, Di Donato N. Baraitser-Winter Cerebrofrontofacial Syndrome. 2015 Nov 19 [updated 2022 Mar 24]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK327153/
70. Baple EL, Salter C, Uhlig H, Wolf NI, Crosby AH. PI4KA-Related Disorder. 2022 Aug 11. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK583038/
71. Fetal akinesia. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 722–727.
72. Jones KL, Jones MC, del Campo M. Adams-Oliver syndrome. In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 444–447.
73. Jones KL, Jones MC, del Campo M. Sturge-Weber syndrome. In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 698–701.
74. Jones KL, Jones MC, del Campo M. Linear sebaceous nevus syndrome. In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 702–705.
75. Ventriculomegaly. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 802–805.
76. Imaging in prenatal diagnosis. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 740–745.
77. Epilepsy in infants and children. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 410–415.
78. Cerebral palsy. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 96–99.
79. Wilmshurst JM, Gaillard WD, Vinayan KP, Tsuchida TN, Plouin P, Van Bogaert P, et al. Summary of recommendations for the management of infantile seizures: Task Force Report for the ILAE Commission of Pediatrics. Epilepsia. 2015;56(8):1185-1197. doi:10.1111/epi.13057.
80. Lundbeck Pharmaceuticals LLC. SABRIL (vigabatrin) tablets and powder for oral solution. Prescribing information. NDA 020427/022006. Effective 2021-10-20. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=a88ac1b4-e2c9-45c0-b321-4785902172e3
81. Seizures with developmental delay/intellectual disability. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 308–311.
82. Galea B, Reid J, Gooley S, Witkowski T, Lane T, Macdonald S, et al. Deep tissue sequencing improves genetic diagnostic yield in focal cortical dysplasia. Prog Neuropsychopharmacol Biol Psychiatry. 2026;149:111920. doi:10.1016/j.pnpbp.2026.111920.
83. García-Cazorla Á, Dionisi-Vici C, Saudubray JM. Disorders of cellular trafficking. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 44. doi:10.1007/978-3-662-63123-2_44.
84. Holoprosencephaly (HPE). In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 168–171.
85. Cerebellar anomalies. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 92–95.
86. Kwan P, Arzimanoglou A, Berg AT, Brodie MJ, Allen Hauser W, Mathern G, et al. Definition of drug resistant epilepsy: consensus proposal by the ad hoc Task Force of the ILAE Commission on Therapeutic Strategies. Epilepsia. 2010;51(6):1069-1077. doi:10.1111/j.1528-1167.2009.02397.x.
87. Jayakar P, Gaillard WD, Tripathi M, Libenson MH, Mathern GW, Cross JH, et al. Diagnostic test utilization in evaluation for resective epilepsy surgery in children. Epilepsia. 2014;55(4):507-518. doi:10.1111/epi.12544.
88. Manickam K, McClain MR, Demmer LA, Biswas S, Kearney HM, Malinowski J, et al. Exome and genome sequencing for pediatric patients with congenital anomalies or intellectual disability: an evidence-based clinical guideline of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2021;23(11):2029-2037. doi:10.1038/s41436-021-01242-6.
89. Rodan LH, Stoler J, Chen E, Geleske T, Council on Genetics. Genetic Evaluation of the Child With Intellectual Disability or Global Developmental Delay: Clinical Report. Pediatrics. 2025;156(1):e2025072219. doi:10.1542/peds.2025-072219.
90. Zuberi SM, Wirrell E, Yozawitz E, Wilmshurst JM, Specchio N, Riney K, et al. ILAE classification and definition of epilepsy syndromes with onset in neonates and infants: Position statement by the ILAE Task Force on Nosology and Definitions. Epilepsia. 2022;63(6):1349-1397. doi:10.1111/epi.17239.
91. Go CY, Mackay MT, Weiss SK, Stephens D, Adams-Webber T, Ashwal S, et al. Evidence-based guideline update: medical treatment of infantile spasms [RETIRED]. Report of the Guideline Development Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology. 2012;78(24):1974-1980. doi:10.1212/wnl.0b013e318259e2cf.
92. European Medicines Agency. Kigabeq (vigabatrin) soluble tablets: EPAR — product information. Available from: https://www.ema.europa.eu/en/documents/product-information/kigabeq-epar-product-information_en.pdf
93. Keenova ARD LLC. ACTHAR GEL (repository corticotropin injection). Prescribing information. Effective 2026-10-13. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=7b48ddec-e815-45f4-9ca0-5c0daaf56f30
94. Khoury J, Blümcke I, Busch RM, Krishnan B, Bulacio J, Bingaman W, et al. Deep characterization of refractory epilepsy due to mild malformation of cortical development with oligodendroglial hyperplasia (MOGHE) and insights into the role of invasive monitoring. Epilepsia. 2026;67(5):2227-2240. doi:10.1002/epi.70118.
95. European Medicines Agency. Votubia (everolimus): EPAR — product information (Summary of Product Characteristics). Available from: https://www.ema.europa.eu/en/documents/product-information/votubia-epar-product-information_en.pdf
96. Moloney PB, Kearney H, Benson KA, Costello DJ, Cavalleri GL, Gorman KM, et al. Everolimus precision therapy for the GATOR1-related epilepsies: A case series. Eur J Neurol. 2023;30(10):3341-3346. doi:10.1111/ene.15975.
97. Carapancea E, Eklund EA, Verhelst H, Cilio MR. Everolimus precision therapy in NPRL2- and NPRL3-related epilepsy. Epilepsia. 2025;66(9):e219-e225. doi:10.1111/epi.18543.
98. ClinicalTrials.gov. NCT03198949: everolimus in patients with refractory seizures who have FCD type II (randomized, double-blind, placebo-controlled crossover). Sponsor: Yonsei University. Available from: https://clinicaltrials.gov/study/NCT03198949
99. Mylan Pharmaceuticals Inc. Sirolimus tablets. Prescribing information (generic of Rapamune, NDA 021083). Effective 2026-03-26. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=5908cd1a-fc5a-462f-99ed-1d8983e253c9
100. European Medicines Agency. Rapamune (sirolimus): EPAR. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/rapamune
101. Jazz Pharmaceuticals, Inc. EPIDIOLEX (cannabidiol) oral solution. Prescribing information. NDA 210365. Effective 2026-05-29. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=8bf27097-4870-43fb-94f0-f3d0871d1eec
102. European Medicines Agency. Epidyolex (cannabidiol): EPAR — product information. Available from: https://www.ema.europa.eu/en/documents/product-information/epidyolex-epar-product-information_en.pdf
103. SK Life Science, Inc. XCOPRI (cenobamate tablets). Prescribing information. NDA 212839. Effective 2025-08-25. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=565c2126-57ae-4e29-b443-723bbe7e2072
104. European Medicines Agency. Ontozry (cenobamate): EPAR — product information. Available from: https://www.ema.europa.eu/en/documents/product-information/ontozry-epar-product-information_en.pdf
105. Liu Y, Daniel EJP, Zhang W, Edmondson A, He M. Longitudinal analysis of clinical and biochemical outcomes during seven years of galactose supplementation in a patient with SLC35A2-CDG. Mol Genet Metab Rep. 2026;49:101375. doi:10.1016/j.ymgmr.2026.101375.
106. ClinicalTrials.gov. NCT05402384: D-galactose (AVTX-801) in SLC35A2-CDG. Available from: https://clinicaltrials.gov/study/NCT05402384
107. Novartis Pharmaceuticals Corporation. VIJOICE (alpelisib) tablets, for oral use; VIJOICE (alpelisib) oral granules. Prescribing information. Revised 7/2025 (SPL set ID 638069a4-aee0-46c9-9e21-54ad449c163c, version 10). Drugs@FDA NDA 215039 (tablets; accelerated approval 05/04/2022) and NDA 218466 (granules; 24/04/2024). https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=638069a4-aee0-46c9-9e21-54ad449c163c
108. U.S. Food and Drug Administration. Drugs@FDA: AFINITOR DISPERZ, NDA 203985 — submission history (ORIG-1 2012-08-29; SUPPL-13 efficacy 2018-04-10). Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=203985
109. European Medicines Agency. Votubia (everolimus): procedural steps taken and scientific information after the authorisation (EMEA/H/C/002311; variations II/0041 and II/0061) [archive]. Available from: https://www.ema.europa.eu/en/documents/procedural-steps-after/votubia-epar-procedural-steps-taken-scientific-information-after-authorisation-archived_en.pdf [accessed 26 October 2026].
110. U.S. Food and Drug Administration. Drugs@FDA: RAPAMUNE, NDA 021083 — submission history (ORIG-1 1999-09-15; SUPPL-55 efficacy 2015-05-28). Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=021083
111. European Medicines Agency. Epidyolex (cannabidiol): procedural steps taken and scientific information after the authorisation (variation II/0005) [archive]. Available from: https://www.ema.europa.eu/en/documents/procedural-steps-after/epidyolex-epar-procedural-steps-taken-scientific-information-after-authorisation-archive_en.pdf [accessed 26 October 2026].
112. European Medicines Agency. Vijoice (alpelisib): EPAR. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/vijoice
113. ClinicalTrials.gov. NCT02451696: pilot study of everolimus on brain mTOR activity and cortical hyperexcitability in TSC and FCD. Sponsor: NYU Langone Health. Available from: https://clinicaltrials.gov/study/NCT02451696
114. ClinicalTrials.gov. NCT04833322: galactose supplementation for MOGHE — pilot trial. Sponsor: Hospital Ruber Internacional. Available from: https://clinicaltrials.gov/study/NCT04833322
115. ClinicalTrials.gov. NCT07287202: An open-label phase 1b/2a study to evaluate the safety and tolerability of oral SVG103 (paxalisib) in adults with FCD-II, TSC or HME. Sponsor: Sovargen. Available from: https://clinicaltrials.gov/study/NCT07287202
116. Kazia Therapeutics. Kazia Therapeutics announces Sovargen Co. doses first patient in phase 1b/2a trial of paxalisib for rare seizure disorders [press release]. 2026 Sep 22. Available from: https://technode.global/prnasia/kazia-therapeutics-announces-sovargen-co-doses-first-patient-in-phase-1b-2a-trial-of-paxalisib-for-rare-seizure-disorders/
117. ClinicalTrials.gov. NCT06392009 (Astroscape, RAD-GRIN-201): radiprodil in patients with TSC or FCD type II. Sponsor: GRIN Therapeutics. Available from: https://clinicaltrials.gov/study/NCT06392009
118. ClinicalTrials.gov. NCT07657975: fluorescence guided focal cortical dysplasia surgery (5-ALA). Sponsor: Hospices Civils de Lyon. Available from: https://clinicaltrials.gov/study/NCT07657975
119. ClinicalTrials.gov. NCT05577754: alpelisib in MCAP (phase II). Sponsor: CHU Dijon. Available from: https://clinicaltrials.gov/study/NCT05577754
120. Stone K, Prinzing G, Lai A, Smith L, Sheidley BR, Corliss MM, et al. Clinical deep sequencing to diagnose pathogenic mosaic variants in malformations of cortical development and epilepsy [preprint]. medRxiv. 2026. doi:10.64898/2026.09.01.26361943.
121. European Medicines Agency. Rapamune (sirolimus): EPAR — product information (Summary of Product Characteristics). Available from: https://www.ema.europa.eu/en/documents/product-information/rapamune-epar-product-information_en.pdf [accessed 26 October 2026].