Lissencephalies and Miller–Dieker Syndrome (17p13.3 Deletion)

Agyria–pachygyria–subcortical band heterotopia spectrum caused by failed neuronal migration, with monogenic causes (PAFAH1B1, DCX, tubulins, ARX, RELN, VLDLR and other genes) and a contiguous gene syndrome, Miller–Dieker syndrome, caused by a 17p13.3 deletion including PAFAH1B1 and YWHAE.

Literature and regulatory information updated: 9 October 2026.

0. Initial Reference Table

1. Introduction

1.1. From clinical descriptions to cytogenetics

The term lissencephaly, “smooth brain”, designates a genetically heterogeneous group of malformations in which the cerebral surface wholly or partly lacks normal gyri and sulci, because postmitotic neurons fail to complete migration from the ventricular zone to the cortical plate [1, 16]. Modern clinical history begins in 1963, when J. Q. Miller described lissencephaly in two siblings in Neurology [43]. Dieker and colleagues published a similar picture with facial dysmorphism in 1969, and Jones and colleagues expanded the phenotype in 1980 and proposed the name “Miller–Dieker syndrome” to distinguish the syndromic form from other lissencephalies, according to the synthesis in Smith's Recognizable Patterns of Human Malformation [34]. Sibling recurrence led to the syndrome being considered autosomal recessive for years. In 1983, Dobyns and colleagues found chromosome 17 abnormalities in two of three unrelated patients (a ring chromosome 17 and an unbalanced translocation with partial 17p13 monosomy) and, reviewing five additional patients from three families with 17p abnormalities, proposed that distal 17p monosomy causes the syndrome [44]. The 1991 series, comprising 27 patients from 25 families, showed severe lissencephaly with a grossly normal cerebellum and characteristic facies in all, a 17p13 deletion visible on karyotyping in 14 of 25 probands and, adding molecular methods, deletion in 21 of 25; of 11 de novo deletions with determined origin, 7 were paternal and 4 maternal [35]. The conclusion was that deletion of a “critical region” with at least two loci at 17p13.3 causes MDS [35].

1.2. Molecular discovery milestones

In 1993, Reiner and colleagues cloned LIS-1 at 17p13.3: approximately 15% of patients with isolated lissencephaly and more than 90% with MDS had microdeletions in a 350 kb critical region, and two non-overlapping deletions of the gene's 5′ and 3′ ends indicated haploinsufficiency as the mechanism; the protein had WD repeats resembling the β subunits of heterotrimeric G proteins [45]. In the same year, in a series of 90 probands (25 MDS and 65 ILS), 17p13.3 deletion was found in 23 of 25 MDS (92%) and 8 of 21 (38%) classical ILS cases, leading to the recommendation of FISH in all patients with classical lissencephaly [46]. In 1994, Hattori and colleagues showed that the gene encodes a subunit of brain platelet-activating factor acetylhydrolase [47]. In 1998, des Portes and Gleeson, in simultaneous studies, identified doublecortin (DCX), an approximately 40 kDa protein highly expressed in fetal neurons, as the cause of X-linked subcortical laminar heterotopia syndrome in females and lissencephaly in males; one of Gleeson's patients had a balanced translocation disrupting the gene [48, 49]. Also in 1998, mice with graded Pafah1b1 reduction demonstrated embryonic lethality in null animals and a dose-dependent migration defect in heterozygotes [50]. Subsequent discoveries were autosomal recessive lissencephaly with cerebellar hypoplasia caused by RELN (2000), XLAG caused by ARX loss of function (2002), the role of YWHAE (14-3-3ε) in exacerbating MDS (2003), narrowing of the critical region distinguishing ILS from MDS to 400 kb (2003), the VLDLR founder deletion in Hutterites (2005) and α-tubulin (TUBA1A) mutations in 2007 [51–56]. In 2021, the review by Koenig, Dobyns and Di Donato counted 31 lissencephaly-associated genes and an overall diagnostic yield above 80% [36].

1.3. Nomenclature and reclassifications

The Barkovich classification, whose latest version is from 2012, groups approximately 200 entities into malformations caused by impaired proliferation or apoptosis, neuronal migration (heterotopia, lissencephaly, cobblestone malformation) and postmigrational organisation (polymicrogyria, schizencephaly, focal cortical dysplasia) [1, 2]. The European Neuro-MIG network consensus (action CA16118 of the European Cooperation in Science and Technology programme, COST), published in 2020 after meetings in Malta, Lisbon and Rehovot and a Delphi process involving 42 experts, established current terminology; the definitions paper had an erratum published the same year [1, 2, 57]. The division between “type I” (classical) and “type II” (cobblestone) lissencephaly, introduced by Dobyns in 1985, was abandoned because cobblestone malformation results from overmigration through pial membrane defects and does not produce a smooth surface; the distinction between “classical lissencephaly” (very thick cortex, two or four layers) and “variant lissencephaly” (slightly thickened cortex, three layers) is also no longer considered valid [1]. Lissencephaly is now classified by degree and gradient of the gyral abnormality, cortical thickness and associated brain malformations, following the 21-pattern imaging classification of Di Donato and colleagues (2017) [1, 58]. The former Dobyns and Kato grading system, 1 to 6, in which grade 1 corresponds to diffuse agyria and grade 6 to SBH, is still used in published genotype–phenotype correlations, but classified only approximately 70% of patients in recent cohorts [58]. The two systems are not equivalent: in Di Donato's clinical scale, grade 3 is most severe, whereas in Dobyns's system grade 1 is most severe, so numbers cannot be transposed between them [3, 58]. The Brazilian book Síndromes Genéticas refers to “type 1 lissencephaly (complete agyria) or 2 (diffuse agyria)” in MDS, wording that confuses types with grades and does not correspond to current nomenclature [59].

1.4. Definition of the entities covered

Agyria is absence of gyri, with sulci more than 3 cm apart, and pachygyria comprises few broad gyri with sulci 1.5 to 3 cm apart; the consensus recommends reserving “agyria” for brains without any detectable sulcus [1, 58]. SBH is a grey matter band separated from the cortex and ventricles by white matter; it belongs to the lissencephaly spectrum because of aetiological overlap and may coexist with pachygyria (for example, frontal pachygyria with posterior SBH) [1, 2]. Microlissencephaly is a separate subgroup combining lissencephaly with severe congenital microcephaly, with head circumference (HC) at birth at or below −3 standard deviations (SD) [2]. PAFAH1B1-associated lissencephaly encompasses ILS, namely lissencephaly as the sole abnormality after careful assessment, and posterior-predominant SBH; the corresponding GeneReview excludes MDS, defined as a larger 17p13.3 deletion of approximately 1.3 Mb including at least PAFAH1B1 and YWHAE [3, 16]. DCX-related disorders encompass anterior-predominant lissencephaly, usually in males, and SBH, mainly in females, and have historical synonyms XLIS, LISX, ILSX, X-linked subcortical laminar heterotopia (X-SCLH), SBHX and “double cortex syndrome” [4]. MDS is a contiguous gene syndrome with severe lissencephaly, facial dysmorphism and variable other organ malformations [3, 35]. Tubulinopathies constitute a spectrum of cortical malformations (lissencephaly, microlissencephaly, dysgyria, simplified gyral pattern) with malformations of the basal ganglia, corpus callosum, cerebellum and brainstem [5]. XLAG combines temporal-predominant lissencephaly with a 6–7 mm cortex, corpus callosum agenesis, neonatal epilepsy, poor temperature regulation, chronic diarrhoea and ambiguous genitalia in 46,XY individuals [3, 16]. RELN-related LCH (Norman–Roberts type) and VLDLR cerebellar hypoplasia (CAMRQ1, cerebellar ataxia with intellectual disability and disequilibrium) belong to the reelin pathway and are distinguished by cerebellar hypoplasia and anterior-predominant “thin” lissencephaly [6, 15, 58].

LIS9 caused by MACF1 is defined by the combination of lissencephaly, almost always posterior-predominant diffuse pachygyria, with a brainstem malformation due to a midline crossing defect: pons and medulla narrow in the dorsoventral axis and broad in the transverse axis, with a median ventral cleft and absent or very scarce transverse pontine fibres, thin anterior commissure and dysplastic hippocampi [8, 24]. The authors of the founding series considered the pattern probably pathognomonic of GAR domain variants, and the 2025 cohort extended it to variants in the adjacent EF-hand domains, proposing the dyadic designation “MACF1-related LISDD” [8, 9]. Irahara syndrome, described in 2014 as a pontine malformation with non-decussating pyramidal tracts and regional polymicrogyria, is the first published patient with this entity and belongs to the brainstem malformations [8, 60].

LIS10 caused by CEP85L is defined by posterior-predominant lissencephaly, from agyria to SBH, with an abrupt transition to near-normal cortex anterior to the central sulcus, without the brainstem malformation of LIS9; OMIM separates it as a distinct entity, and Orphanet groups it with LIS9 under ORPHA:572013 [11, 12, 24, 61]. The phenotype is milder than that of posterior PAFAH1B1 lissencephaly, to the extent of allowing vertical transmission in several families [12].

2. Molecular Genetics and Pathophysiology

2.1. Neuronal migration and origin of the spectrum

In the dorsal telencephalon, neuronal proliferation begins at approximately the 5th–6th gestational week, decreases after 16 weeks and ends between 22 and 25 weeks; migration begins at the same time, peaks between the 3rd and 5th months and ends between 30 and 35 weeks, continuing in the cerebellum until the middle of the second year of life [1]. Most cortical neurons undergo radial migration guided by apical radial glia, changing from bipolar to multipolar in the intermediate zone; GABAergic interneurons undergo tangential migration from the medial and caudal ganglionic eminences [1]. When the molecular machinery of movement fails, namely tubulins, microtubule-associated proteins (MAPs) such as LIS1 and DCX, and actins, neurons stop before reaching the cortical plate and the lissencephaly–microlissencephaly–dysgyria spectrum emerges; when the pial membrane fails, neurons overshoot their destination, producing polymicrogyria and cobblestone malformation [1]. In thick classical lissencephaly, the cortex has four layers: molecular, superficial cellular (pyramidal), cell-sparse and deep cellular, the latter formed by arrested neurons, corresponding to the three strata resolved on MRI [1, 58]. Neuropathology distinguishes the PAFAH1B1 pattern (four layers with a posterior gradient, relatively thick layer 2) from the DCX pattern (anterior gradient, thin layer 2 transitioning to small subcortical heterotopia nodules) and the two-layer lissencephaly of severe tubulinopathies [58].

2.2. PAFAH1B1 (LIS1)

The PAFAH1B1 gene lies at 17p13.3 and has the official name platelet activating factor acetylhydrolase 1b regulatory subunit 1 (synonyms LIS1, MDCR, NudF); its MANE Select transcript (Matched Annotation from NCBI and EMBL-EBI, a reference transcript agreed between NCBI and EMBL-EBI) is NM_000430.4 (NP_000421.1) [13, 14]. The literature alternates between “alpha subunit” and “beta subunit”, and the GeneReview itself uses both, so the official HUGO Gene Nomenclature Committee (HGNC) designation is adopted [3, 14]. The highly conserved protein has an N-terminal LisH motif and a coiled-coil region responsible for dimerisation, followed by seven C-terminal WD40 repeats mediating protein–protein interactions [3]. With PAFAH1B2 and PAFAH1B3, it forms the trimeric complex regulating platelet-activating factor in the brain and interacts directly with tubulins and indirectly with MAPs such as dynein, DCX and NDE1, participating in cytoskeletal organisation, mitotic spindle orientation, interkinetic motility, nucleokinesis and radial and tangential migration [3]. LIS1 positively regulates cytoplasmic dynein, the motor directed towards microtubule minus ends: the neuron extends a leading process and dynein at the nuclear surface translocates the nucleus towards the centrosome [15]. The disease mechanism is loss of function through haploinsufficiency, demonstrated by deletions and mice: homozygous null animals die shortly after implantation, heterozygotes have cortical, hippocampal and olfactory bulb disorganisation due to cell-autonomous migration delay, and further activity reductions worsen the phenotype [45, 50]. Two alternative transcripts of 5.5 and 7.5 kb differ in the 3′ untranslated region [3]. There is no biochemical disease marker [15].

Genotype–phenotype correlation is based on the former Dobyns grading system: telomeric deletions including the gene's 5′ end are mainly associated with grades 2–3, intragenic deletions or duplications with grade 3, 3′-end deletions with grades 3–4 and intragenic variants with grades 3 or 4 [3]. Variants predicting premature termination cause more severe phenotypes than missense variants, and truncating variants early in the gene (coiled-coil domain) may be more severe than distal ones, with exceptions described in both directions [3, 62]. In the cohort of 811 patients, PAFAH1B1 variants were almost always loss-of-function (21 of 23 on the panel) and the phenotype was almost always anterior pachygyria with posterior agyria [7]. Postzygotic mosaicism attenuates the phenotype: PAFAH1B1 appears in posterior, never diffuse, SBH, generally with an intermediate allele fraction of approximately 30%, and the same variant NM_000430.4:c.817C>T p.(Arg273Ter), pathogenic in the public ClinVar database (VCV000008074), was described in mosaic form in partial posterior SBH and constitutionally in severe lissencephaly [7, 13, 63]. The GeneReview brings together nine individuals with posterior-predominant SBH caused by missense, frameshift variants and somatic mosaicism for missense, nonsense, in-frame and splicing variants, and somatic mosaicism for PAFAH1B1 deletion is associated with mild to severe delay [3]. By 2021, nearly 200 individuals with an intragenic PAFAH1B1 variant had been reported [3]. No other phenotypes are known for germline intragenic variants; 17p13.3 duplications including the gene constitute a distinct allelic entity (section 6) [3].

2.3. The 17p13.3 region and Miller–Dieker syndrome

The typical MDS deletion is terminal and encompasses PAFAH1B1, YWHAE and frequently other contiguous genes; the GeneReview describes an approximately 1.3 Mb region, a 2025 review cites approximately 3.4 Mb (from YWHAE to ANKFY1) and a 2026 report mentions 26 genes, figures reflecting breakpoint variability rather than a fixed size [3, 64, 65]. The 2003 map, using FISH in 19 ILS, 11 MDS and 4 deletions without LIS1, narrowed the region distinguishing ILS from MDS to 400 kb and identified 8 consistently deleted genes in MDS; combined CRK and YWHAE deletion delineated patients with the most severe lissencephaly [54]. YWHAE encodes 14-3-3ε, which binds NUDEL phosphorylated by CDK5/p35 and maintains that phosphorylation; Ywhae-deficient mice have migration defects and double Pafah1b1/Ywhae heterozygotes have more severe defects, explaining greater lissencephaly severity in MDS [53]. Isolated or syndromic lissencephaly results from PAFAH1B1 haploinsufficiency: ClinGen (Clinical Genome Resource) assigns it a haploinsufficiency score of 3 (sufficient evidence), while YWHAE scores 1 (little evidence) [15]. YWHAE is loss-of-function intolerant (probability of loss-of-function intolerance, pLI, 0.98) and proximal CRK is predicted to be haploinsufficient and has been associated with postnatal growth restriction [66, 67]. The deletion may arise as a simple terminal deletion, a ring chromosome 17, an unbalanced translocation inherited from a balanced reciprocal translocation carrier or a recombinant chromosome 17 through crossing-over in a pericentric inversion carrier [34, 44]. The most severe cases have cytogenetically visible deletions and unbalanced rearrangements with duplication of another chromosome segment [3]. Small telomeric deletions of 600 kb or less have been described in normal individuals, requiring interpretation of each deletion's gene content rather than cytoband alone [68].

2.4. DCX

The DCX gene (Gene ID 1641) is located at Xq23 according to NCBI Gene and the GeneReview; the Oxford Desk Reference location Xq22.3 and the designation Xq22.3-q23 in the 1998 articles reflect contemporary mapping [4, 14, 16, 49]. The current reference and MANE Select transcript is NM_001195553.2 (NP_001182482.1); the GeneReview writes “NM_0011955553.2”, a typographical error with an extra digit [4, 13]. Older publications use cDNA NM_178153, which has 15 fewer bp at the end of exon 5 and 3 fewer bp at the end of exon 6, shifting c. positions without altering the reading frame; historical variants must be converted before comparison with current reports [4]. Doublecortin is a MAP with two tandem DC domains (N-DC and C-DC) absent from other MAPs; it promotes microtubule polymerisation and stabilisation, concentrates at neuronal process tips and axonal regions generating collaterals, interacts with spinophilin and F-actin and participates in embryonic neuromuscular junction formation [4, 7]. Functional studies indicate loss of function through different cellular mechanisms depending on the variant [4].

In hemizygous males, all neurons express the variant and the cortex becomes thick, smooth and disorganised, with an anterior gradient; in heterozygous females, X-inactivation creates two populations, one expressing the normal allele and completing migration and another expressing the variant and stopping in white matter, forming the heterotopic band [4, 16]. This inheritance is described as X-linked semidominant in the Oxford Desk Reference, X-linked in the GeneReview and X-linked recessive in Orphanet; all three designations refer to the same fact, frequent but variable expression in heterozygous females [4, 16, 19]. X-inactivation pattern analysis has no proven diagnostic or predictive value [4]. Approximately one-third of DCX variants recur, with similar cortical phenotypes within and between families [4, 69]. Constitutional hemizygous nonsense variants and whole-gene deletions are extremely rare in males, suggesting that complete loss of function is probably lethal; most truncating variants in males result from postzygotic mosaic events [4, 70]. Early truncating variants in males produce near-complete agyria, difficult to distinguish from a PAFAH1B1-YWHAE deletion or a severe TUBA1A variant [4, 7]. Hemizygous missense variants in the N-DC domain tend to cause more severe lissencephaly than those in the C-DC domain; truncating variants are more associated with generalised bands and missense variants with frontal bands; loss-of-function variants predominate in sporadic cases and missense variants in familial cases [4]. Asymptomatic heterozygous females more frequently have missense variants at codon p.Arg196 (published designation, without a c. description) [4].

The 2025 GeneReview lists the most cited DCX variants; VariantValidator normalisation confirmed the c. description on the MANE transcript for variants listed in the following table [4, 13].

Together, hot spots account for more than one-third of pathogenic sequence variants; the GeneReview states that variants were supplied by the authors and their classification was not verified by the editors [4, 69]. Codon p.Arg196 appears without a c. description in the source.

2.5. Tubulinopathies

Tubulin genes associated with cortical malformations are TUBA1A (α-1A chain), TUBB (formerly TUBB5), TUBB2A, TUBB2B, TUBB3, TUBG1 (γ-1) and, with recessive inheritance, TUBA8 [5, 16]. TUBA1A (12q13.12; the Oxford Desk Reference indicates 12q12-14) encodes the predominant α isotype in differentiating neurons; ClinGen assigns haploinsufficiency and triplosensitivity scores of 0, consistent with a dominant-negative or gain-of-function effect of missense variants [2, 15]. Many variants are private and difficult to classify without functional studies, and the typical imaging pattern helps interpret variants of uncertain significance [2].

The table summarises Table 1 of the tubulinopathy GeneReview [5]. In the cohort of 811 patients, TUBA1A explained 38 cases, distributed among LCH with congenital microcephaly (14%), LCH with normal or borderline HC (35%), dysgyria (29%) and PAFAH1B1-like posterior agyria-pachygyria (14%), the latter always at codon 402 [7]. Variants NM_006009.4:c.1204C>T p.(Arg402Cys) and c.1205G>A p.(Arg402His) are classified in ClinVar as pathogenic/likely pathogenic (VCV000160146) and pathogenic (VCV000007071), respectively; the former produces posterior agyria-pachygyria nearly indistinguishable from that of PAFAH1B1 and the latter diffuse agyria with mild vermian hypoplasia [3, 13, 58, 63]. TUBB2B accounted for 1.2% of the cohort, mainly with dysgyria, and TUBG1 for 0.7%, with recurrent p.Ser259Leu in five of six patients and a mild PAFAH1B1-like posterior pattern [7].

2.6. ARX

ARX (aristaless related homeobox, Xp21.3; NM_139058.3) encodes a transcription factor required for proliferation, tangential migration and differentiation of GABAergic interneurons, making XLAG an “interneuronopathy”; it also specifies pancreatic α cells and enteroendocrine lineages, explaining chronic diarrhoea [13, 15]. Mutant male mice have small brains, abnormal interneuron migration and abnormal testicular differentiation, and ARX loss of function was found in individuals with XLAG and some female relatives [52]. The allelic spectrum is continuous and genotype predicts phenotype: truncating variants (large deletions, frameshift, nonsense and splicing variants in exons 1 to 4) cause XLAG or hydranencephaly with abnormal genitalia; non-conservative homeodomain missense variants cause less severe XLAG; a conservative homeodomain substitution causes Proud syndrome (corpus callosum agenesis with abnormal genitalia) [15, 71]. At the mild end, with a structurally normal brain, polyalanine tract expansions and duplications, most missense variants and exon 5 deletion cause X-linked infantile spasms, developmental and epileptic encephalopathy type 1 (DEE1, Ohtahara pattern), X-linked myoclonic epilepsy, Partington syndrome (intellectual disability with episodic hand dystonia) and nonsyndromic X-linked intellectual disability [15]. More than half of reported mild-spectrum cases have the recurrent 24 bp duplication in the second polyalanine tract, historically described as c.429_452dup and normalised by the 3′ rule as NM_139058.3:c.441_464dup p.(Ala148_Ala155dup) [13, 15]. These allelic forms without lissencephaly lie outside this article's core and are mentioned as differentials and allelic disorders [15].

2.7. Reelin pathway: RELN and VLDLR

Reelin is an extracellular matrix glycoprotein secreted by Cajal–Retzius cells during development; it binds two low-density lipoprotein receptor family receptors, VLDLR and ApoER2 (LRP8), and induces DAB1 phosphorylation, a cascade enabling completion of migration in the cortex and cerebellum [6, 15]. Biallelic RELN variants (7q22.1; NM_005045.4) cause lissencephaly 2, an LCH with hippocampal and brainstem malformation; the original 2000 description involved two splicing mutations with low or undetectable reelin [15, 51]. In the cohort of 811 patients, RELN explained seven cases, including compound heterozygosity with a single-exon deletion, so RELN testing must include copy-number analysis; a boy homozygous for NM_005045.4:c.1249C>T p.(Gln417Ter), described in the original article using transcript NM_173054.2, had a more severe phenotype than previous patients and died in the second decade [7, 13]. Heterozygous RELN variants cause an allelic disease without malformation, autosomal dominant epilepsy with auditory features (familial temporal lobe epilepsy 7), beginning between 10 and 30 years with penetrance of approximately 60% [15]. VLDLR (9p24.2; NM_003383.5) encodes an 873-amino-acid receptor expressed in the heart, muscle, kidney and brain; all known pathogenic variants are predicted to cause loss of function [6]. In Hutterites, a homozygous 199 kb deletion encompassing the entire gene constitutes a founder effect, described in eight patients from three related families, the first malformation syndrome attributed to a lipoprotein receptor [55]. Variant NM_003383.5:c.1256G>A p.(Cys419Tyr), in the EGF-B domain, is associated with a less severe phenotype [6, 13]. Genomic coordinates of the Hutterite deletion were not normalised to a reference build in the consulted sources.

2.8. Genetic architecture and other genes

Lissencephaly is now considered an exclusively genetic disease [2]. In the largest published series, 811 patients from the Seattle and Florence cohorts, 655 (81%) received a molecular diagnosis; PAFAH1B1 explained 322 cases (40%: 75 PAFAH1B1-YWHAE deletions and 247 PAFAH1B1 alterations), DCX 187 (23%: 73 males and 114 females), TUBA1A 38 (5%) and DYNC1H1 23 (3%), and these four genes explained 70.5% of the cohort (69% in the article abstract) [7]. ARX explained 31 cases, a frequency inflated by active recruitment and probably below 1% in reality; next were TUBB2B (10), ACTG1 (9), ACTB (8), RELN (7), TUBG1 (6), CRADD (6), TUBB (4), TUBB3 (2) and VLDLR (2) [7]. The authors proposed the first classification by biological networks, with eight groups reflected in imaging.

The table follows Table 2 of Di Donato and colleagues (2018) [7]. DYNC1H1 had 22 missense variants and one possible splicing variant, all de novo when both parents were tested (13 patients), with recurrent variants NM_001376.5:c.926G>A p.(Arg309His) and NM_001376.5:c.10031G>A p.(Arg3344Gln) [7, 13]. Recessive CRADD reduces neuronal apoptosis through failed caspase-2 activation and accounts for less than 1% of all lissencephalies but one-third of thin lissencephaly with a normal cerebellum [7]. Other described genes include CDK5 (recessive, early lethal), NDE1, KATNB1 and WDR62 (recessive microlissencephaly), KIF2A, KIF5C, APC2 (recessive), RNU4ATAC (microcephalic osteodysplastic primordial dwarfism type I), CEP85L (posterior-predominant lissencephaly 10) and PIDD1 (recessive) [3, 4, 36, 58, 72]. In 2026, a Chinese cohort first associated DARS2 and NPRL3 with lissencephaly, based on isolated cases [73].

2.9. Penetrance, expressivity and modifiers

Except for X-linked genes (ARX, DCX) and GATOR1 genes, causal cortical malformation variants appear fully penetrant structurally, although carriers may be clinically asymptomatic; parental MRI is therefore essential when a variant is inherited from an apparently unaffected parent [2]. No asymptomatic carrier has ever been described among males with a hemizygous germline DCX variant, whereas postzygotic mosaics may have mild presentations [4]. In females, DCX penetrance was below 50% in heterozygous or mosaic mothers of affected children, and 14 of 25 heterozygous mothers (56%) were asymptomatic with normal MRI [4, 69]. Monozygotic twins heterozygous for the same nonsense variant both had generalised thick SBH, but with different band thickness, cognitive severity and electroencephalograms (EEGs), showing that X-inactivation and other factors modulate expression [4]. In MDS, YWHAE acts as a lissencephaly severity modifier and CRK as a growth modifier [53, 54, 67]. In cerebral organoids from patients with PAFAH1B1 lissencephaly of different severities, LIS1 protein levels varied between lines and partially correlated with clinical severity, suggesting variant-specific effects [74].

2.10. MACF1 and lissencephaly with decussation defects

MACF1 (1p34.3) encodes microtubule-actin crosslinking factor 1 (also called ACF7), one of the two mammalian spectraplakins (the other is dystonin, encoded by DST), which links actin filaments and microtubules and integrates cytoskeletal processes [8]. The gene expresses many isoforms through alternative splicing, some brain-specific; the N-terminal end of the main isoforms contains two calponin homology (CH) domains that bind actin, a plakin domain and a long rod of spectrin repeats, and the C-terminal end, common to all isoforms, binds microtubules through two calcium-binding EF-hand domains, a zinc-binding GAR domain (growth arrest-specific protein 2, Gas2, related), a Gly-Ser-Arg region and an SxIP EB1-binding motif [8, 9]. The GAR domain forms an α-β sandwich fold, and four conserved residues (three cysteines and one aspartate), invariant across spectraplakins and essential for microtubule binding, coordinate the zinc ion [8].

In the mouse, complete loss of Macf1 is lethal shortly after gastrulation (E10.5–E11.5), with a phenotype similar to that of Wnt3 loss, because MACF1 participates in WNT signalling through axin, GSK3B, APC, LRP5/LRP6 and β-catenin [8]. In the model of Chen and colleagues (2006), Macf1−/− embryos died as early as gastrulation, with developmental delay at E7.5 and defects of the primitive streak, node and mesoderm, which places lethality earlier than the interval cited in 2018 [17]. Deletion restricted to skeletal muscle is compatible with life but disorganises the microtubule network of the fibres, progressively displaces myonuclei and alters the neuromuscular junction, an experimental finding relevant to the neuromuscular presentations described with biallelic variants [17]. Nervous system-targeted deletion reproduces the human malformation: pyramidal neurons and interneurons migrate slowly and unidirectionally, with short and unstable leading processes, centrosomes that fail to advance and unstable microtubules; there are similar defects in the hippocampus, and the anterior and hippocampal commissures are severely hypoplastic, with fewer fibres in the corpus callosum and misrouted thalamocortical tracts [8]. In Drosophila, the orthologue Short stop interacts with Robo and Slit and regulates axonal midline crossing [8, 9]. Phosphorylation of the GSR domain by GSK3B uncouples MACF1 from microtubules and inhibits neuronal migration in the mouse [9].

The domain–phenotype correlation is the central element for interpretation. In the founding 2018 series, seven of eight children had heterozygous de novo missense variants at three of the four zinc-binding residues of the GAR domain, and the eighth (the Irahara patient) had an in-frame 39.6 kb intragenic deletion of exons 62 to 93 of transcript MACF1-204, which removes the last ten spectrin repeats, both EF-hand domains and the first two zinc-binding residues; none of the missense variants was present in gnomAD [8]. A ninth child, with a de novo variant in spectrin repeat SR4 (a region that binds GSK3B, axin and APC), had a similar mild posterior-predominant lissencephaly, but only subtle brainstem dysplasia [8]. The 2025 international cohort, with 45 individuals (32 monoallelic and 13 biallelic), identified LISDD in 10: four with substitutions of previously known zinc-binding residues, two with substitutions of the most N-terminal zinc-binding residue, one with a variant in the β2 strand of the GAR domain that does not affect a zinc residue, two with de novo missense variants in the EF-hand domains and one with a homozygous nonsense variant in exon 1 [9]. The 15 known individuals with substitution of a zinc-binding residue had the same brain and clinical phenotype, reflecting a strong genotype–phenotype correlation [9].

The mechanism of variants in the EF1-EF2-GAR module is a toxic gain of function. The 2018 authors proposed an altered-function or dominant-negative effect, because ExAC recorded 40 heterozygous truncating MACF1 variants (19 stop, 16 frameshift and 5 splice donor) without this phenotype, even though the gene is among the most intolerant to loss of function (pLI = 1.0) [8]. In 2025, EF1-EF2-GAR constructs carrying the variants p.Met7044Thr, p.Asp7095Gly, p.Cys7130Tyr and p.Arg7147Pro (MACF1-204 notation, ENST00000372915.8), expressed in HeLa and HEK293T cells, bound more strongly to bundled and curved microtubules than the normal construct, a mechanism comparable to that of the tubulinopathies and of the microtubule-associated proteins that cause lissencephaly; the p.Cys7130Tyr and p.Arg7147Pro constructs were almost insoluble, indicating marked destabilisation of the domain [9]. Fibroblasts from two patients formed a normal number of cilia, but with a higher proportion of short cilia (<3 µm), a length-control defect similar to that of cells with RTTN variants [8]. For the decussation defect, the 2025 authors hypothesise increased axonal repulsion at the midline through the human orthologue of Kra (BZW1) and ROBO2/SLIT1, without excluding a direct effect of microtubule dysfunction on axon guidance [9].

Nomenclature requires attention, because publications use different transcripts. The 2018 series described the variants on transcript MACF1-204 in an earlier version (ENST00000372915.7) and on NM_012090.5, which is shorter and lacks the plectin repeats; the 2025 cohort used ENST00000372915.8, whose GAR residue numbering is three units lower (2018 p.Asp7186Tyr corresponds to 2025 p.Asp7183Tyr), and gave in a supplement the notation on the MANE Select transcript NM_001394062.1 (ENST00000564288.6, MACF1-230) [8, 9]. The following table presents the LISDD variants on the MANE transcript, converted and validated in VariantValidator (GRCh38) and cross-checked against ClinVar [13, 63].

The ClinVar classification of p.(Cys7234Tyr) as of uncertain significance rests on a single submission without phenotype and does not incorporate the 2025 clinical, imaging and functional data, which support pathogenicity [9, 63]. The de novo SR4 spectrin repeat variant of the 2018 series (p.Gly6664Arg on ENST00000372915.7) corresponds to NM_012090.5:c.14116G>C p.(Gly4706Arg), the notation recorded by OMIM, and to NM_001394062.1:c.20293G>C p.(Gly6765Arg) on the MANE transcript (VariantValidator, GRCh38); ClinVar classifies it as pathogenic/likely pathogenic for LIS9, without assertion criteria [8, 13, 17, 63]. OMIM counts this patient among the nine with LIS9 (five missense variants and one deletion), whereas the 2025 cohort restricts LISDD to the EF1-EF2-GAR module, a divergence that weighs on the interpretation of variants in the spectrin repeats [9, 10].

The homozygous p.(Arg24Ter) case showed that MACF1 has at least five transcripts with distinct transcription start sites and exons 1 (MACF1-203/204, -226, -230, -234 and -235), all spliced to the same exon 2, where the first CH domain begins [9]. MACF1-226 predominates in fibroblasts; in the fetal frontal cortex (18–21 post-conceptional weeks) and in the adult brain MACF1-204 predominates, and in the fetal cerebellum MACF1-204 and MACF1-226 each contribute about half of expression, which explains why loss of MACF1-204 exon 1 causes anterior-predominant lissencephaly with a normal cerebellum [9].

Outside the EF1-EF2-GAR module, the association of MACF1 with disease remains controversial. In the 2025 cohort, 8 of the 23 heterozygous variants outside this module were considered unlikely causes, and the remaining 15 were associated with phenotypes so heterogeneous that no correlation was possible; among biallelic variants, mainly in the plakin domain, 9 individuals from 8 families had a neurodevelopmental disorder with ID, neuropathy, a neuromuscular presentation or a movement disorder, without confirmation of the association with congenital myasthenic syndrome previously proposed in two patients [9]. In 2025, a Chinese generalised epilepsy cohort identified two heterozygous de novo and eight biallelic variants, and in 2026 hierarchical clustering of the phenotypes of 22 individuals associated biallelic variants outside the GAR domain with craniofacial and skeletal involvement, with results the authors describe as hypothesis-generating [23, 76]. Also in 2026, two de novo truncating variants in the spectrin repeats, one with global delay and the other with focal seizures and borderline intelligence, were interpreted as haploinsufficiency, an interpretation that contrasts with the presence of heterozygous truncations in healthy individuals [8, 9, 77]. De novo MACF1 variants have also been observed in schizophrenia cohorts, and a partial 1p34.3 duplication including more than half of the 3′ end of the gene segregated with hypotonia in one family, associations that do not establish Mendelian disease [8]. In that family, described in 2014, three siblings and their mother, mildly affected and a probable de novo carrier, had hypotonia of variable severity; the two most affected had generalised weakness from birth, fatigability, facial weakness and contractures, and muscle biopsy in two showed reduced MACF1 expression with non-specific ultrastructural changes [10].

2.11. CEP85L and lissencephaly 10

CEP85L (6q22.31) encodes an 805-amino-acid centrosomal protein with a C-terminal coiled-coil domain, which localises to the pericentriolar region and, in migrating cortical neurons of the mouse embryo, to the centrosome [12, 18]. The gene has two full-length transcripts (NM_001042475 and NM_001178035) with different 5′ untranslated regions and first coding exons; NM_001042475, the MANE transcript, predominates in all tissues studied, including the fetal brain [12, 13]. CEP85L binds the mother centriole and forms a complex with proteins required for neuronal migration (CDK5, LIS1, NDE1, KIF2A and DYNC1H1); its loss disrupts the localisation and activation of CDK5, the centrosome and the microtubule cytoskeleton [61]. The shared centrosomal localisation with LIS1 is the proposed explanation for the posterior predominance of both malformations [12, 61]. In the mouse, Cep85l knockdown by in utero electroporation at E13.5 reduced the proportion of neurons reaching the cortical plate at E17.5 from 61.5% to 42.2%, a defect partially rescued by re-expression of the gene and still present on postnatal day 7, without altered neuronal maturation [12].

The mechanism is not simple haploinsufficiency. The gene is tolerant to loss of function and to missense variation (Z = 0.4), and truncating variants occur in healthy individuals, but pathogenic variants cluster in a highly constrained 15-amino-acid region in exon 2 [12]. Exon 2 donor variants (c.232+1, +3 and +5) cause or are predicted to cause skipping of this exon, with an in-frame deletion of 53 residues, p.(Gly25_Glu77del), and start-loss variants use a downstream methionine, p.(Met103), with loss of residues 2 to 103; both remove the same region and maintain an abnormal protein at low levels, which led the authors to propose a dominant-negative effect [11, 12]. In the 2020 series, two missense variants in this region appeared in 54 alleles from patients with SBH, versus 2 in 31,400 gnomAD alleles (p < 0.00001) [12]. In 2025, the 18 known variants in 32 individuals (10 missense, 5 splicing, 2 start-loss and 1 nonsense) all clustered between amino acids 1 and 103 [21].

The variants in the table were validated in VariantValidator (GRCh38, MANE transcript), and the classifications are those of ClinVar in November 2026 [13, 63]. The start-loss variant of the first patient is recorded in OMIM as c.1T>C, a notation incompatible with the reference sequence (the first nucleotide of the ATG codon is an adenine); the 2020 paper, ClinVar and VariantValidator give c.2T>C [12, 13, 18, 63]. The Canadian family was published on NM_001042475.2, with the same position on NM_001042475.3 [13, 79].

3. Clinical Spectrum and Manifestations by System

3.1. Severity scale and relationship with imaging

Lissencephaly clinical severity correlates roughly but consistently with malformation extent and degree. The clinical scale proposed by Di Donato and colleagues in 2017, reproduced in the PAFAH1B1 GeneReview, applies to all causes and is not gene-specific [3, 58].

The authors emphasise that exceptions are frequent, particularly because severe epilepsy can lower overall functioning below that predicted by imaging [58]. Problems weighing most heavily on prognosis and life expectancy are feeding difficulties (gastro-oesophageal reflux and aspiration), epilepsy and pneumonia; between 35% and 85% of children with thick classical lissencephaly develop infantile spasms, often with electrophysiological evolution to Lennox–Gastaut syndrome (LGS), and cognitive development slows with seizure onset [58]. Box 1 of the 2020 consensus summarises the age-specific presentation of malformations of cortical development (MCD): in fetal life, reduced movements, polyhydramnios and ultrasonographic abnormalities; at birth, abnormal HC, dysmorphism, congenital abnormalities, abnormal tone and feeding and respiratory difficulties; in infancy, global delay, hypotonia or hypertonia and epilepsy with spasms; in childhood, cerebral palsy, seizures, speech delay and visual impairment; in adolescence and adulthood, epilepsy and ID [2].

3.2. PAFAH1B1-associated lissencephaly

Neonates are normal or have mild to moderate hypotonia, feeding difficulties and poor head control; during the first years, poor visual and auditory tracking, axial hypotonia and mild distal spasticity develop, sometimes with opisthotonus, and later prominent distal spasticity with persistent axial hypotonia, which may progress to moderate spastic tetraparesis and scoliosis [3]. Seizures occur in more than 90% of patients with ILS, usually within the first two years, with mean onset at 5 months; approximately 80% have infantile spasms, although EEG does not always show typical hypsarrhythmia [3, 39, 62]. Epileptic encephalopathy typically evolves from West syndrome to LGS; after the first months, persistent spasms, focal motor, generalised tonic, atypical absence, atonic and myoclonic seizures coexist, and more than 65% of patients are drug-resistant [3]. Before seizures, delay is mild; spasm onset may accompany functional decline, and even with good control the maximum developmental level in complete lissencephaly is approximately 3 to 5 months: some roll and creep, very rarely sit, and few walk with support [3]. Less severe lissencephaly (Dobyns grade 4, mainly partial posterior disease or pachygyria) has a better course [3]. HC at birth is normal (between the mean and −2 SD) and most become microcephalic by the first year because of slow postnatal head growth [3]. Feeding difficulties improve in the first months, worsen with seizure onset and again with age and size, and at least 50% of children with PAFAH1B1 lissencephaly (but not SBH) eventually require gastrostomy; constipation is frequent [3]. Poor airway control leads to aspiration pneumonia, the most common terminal event [3]. Visual interaction is poor or absent in a subset, probably resulting from cortical malformation [3].

The German–Austrian–Swiss cohort published in 2026 describes 47 patients with MRI- and genetically confirmed lissencephaly, recruited through a Regensburg genetics centre, the Liss e.V. patient association and paediatric neurologists; declared limitations are the retrospective questionnaire design, recruitment bias and no adjustment for multiple comparisons [33]. The following table reproduces the main characteristics of the 38 patients with PAFAH1B1 (median age 6 years) and 5 males with DCX (median 10 years) [33].

In this cohort, among PAFAH1B1 patients, 66.8% acquired head control, 58.8% babbled, 55.6% rolled and 85% fixed their gaze, but unsupported sitting, crawling, standing and supported walking occurred in fewer than 20%, and only two walked independently; one-third never acquired head control [33]. Six of nine reassessed PAFAH1B1 patients regressed (loss of head control, grasping, rolling or visual fixation), associated with worsening epilepsy and nutritional status [33]. Males with DCX achieved more milestones, with statistically significant differences for unsupported sitting, independent walking, pointing, single words, sentences and picture naming [33].

Posterior-predominant PAFAH1B1 SBH, generally mosaic, is associated with mild to severe delay in all reported cases and seizures typically beginning in the first decade, frequently drug-resistant and potentially causing regression [3]. Aspiration pneumonia is not typical of SBH [3].

3.3. DCX-related disorders

In hemizygous males, the spectrum ranges from thick anterior-predominant lissencephaly, with a four-layer cortex and thin layer 2, to rare mixed phenotypes of frontal pachygyria with posterior SBH; by 2025, more than 300 individuals with a pathogenic DCX variant had been reported [4, 7]. Motor development is delayed but generally better than in PAFAH1B1 lissencephaly [4]. The following table brings together data from two cohorts published by Leger (2008) and Bahi-Buisson (2013), presented in the GeneReview [4, 69, 70].

In Leger's series of 33 males (median 7.5 years, range 1.5 to 37), almost half walked independently, all had significant cognitive and language impairment, almost half were nonverbal, 30% had moderate behavioural abnormalities (autism, sleep disturbance, agitation) and 12% extreme irritability; ten were adults [4, 70]. Epilepsy affected approximately 75% of males in that series, with two onset peaks, within the first 6 months (approximately half) or at a median of 6 years (2 to 17), including infantile epileptic spasms syndrome (IESS), and was intractable in more than half; the 83% in the table refers to the combined denominator of 76 males [4]. Early onset correlates with refractoriness [4]. Abnormal tone and limited mobility lead to contractures and scoliosis, and severe forms compromise swallowing [4].

In heterozygous females, SBH is the most common malformation, with variable anterior pachygyria, and phenotype correlates with band extent and thickness [4]. Bahi-Buisson and colleagues distinguished a severe subgroup with a thick band, usually sporadic, and a mild subgroup, mainly asymptomatic heterozygotes with normal MRI or only thin frontal bands [4, 69]. In the severe phenotype, more than 80% have childhood seizures (focal, atypical absences, combined atonic and tonic seizures, spasms), LGS occurs in approximately 55% with thick bands and 50% with thin bands, median age at first seizure is 2.2 years with a thick band and 10 years with a thin band, almost all female probands have moderate to severe ID, 84% of those with thick bands have severe language impairment, limited verbal output or no speech, and approximately 78% have moderate to severe behavioural problems [4]. At the other extreme, women with a thin frontal band may have average or slightly reduced cognition without other symptoms and be recognised only after diagnosis in an offspring [4]. Rare males with mosaicism or missense variants retaining residual function have mild SBH resembling the female phenotype [4]. A family with p.(Lys201Glu) shows the extent of intrafamilial variability: a 78-year-old grandmother and her 50-year-old daughter without seizures, with normal intelligence quotient (IQ) and neuropsychological profile despite frontocentral SBH-pachygyria, and a 35-year-old granddaughter with seizures from 9 years [4].

3.4. Miller–Dieker syndrome

MDS combines severe lissencephaly (grades 1–2 of the historical Dobyns scale, namely diffuse or near-diffuse agyria), characteristic facial dysmorphism, variable malformations and severe neurological impairment [3, 35]. Facies includes a high prominent forehead with vertical ridges and furrows in the central forehead, more visible during crying, bitemporal narrowing, a short nose with anteverted nostrils, a protruding upper lip with thin vermilion and downturned corners, micrognathia, apparently low-set ears and delayed tooth eruption [3, 34, 35]. In Smith, corpus callosum agenesis or hypoplasia occurs in 74%, a large cavum septi pellucidi in 77% and small midline calcifications in 45%, without an explicit denominator; absence of frontal and temporal opercula leaves the Sylvian fissure widely open, with a figure-of-“8” appearance on computed tomography (CT) [34]. Other features are cryptorchidism, a pilonidal sinus, 5th-finger clinodactyly and polyhydramnios; occasional findings include congenital heart disease (tetralogy of Fallot, ventricular septal defect, pulmonary stenosis), intrauterine growth restriction, omphalocele, a pelvic kidney, cystic renal dysplasia, lipomeningocele with tethered cord, cleft palate and cataract [34]. A 2026 perioperative report adds conotruncal heart defects, septal defects, patent ductus arteriosus and anomalous origin of the right coronary artery [65]. HC at birth is typically normal and most become microcephalic by the first year; delay and hypotonia predominate before seizures, and smiling, brief visual fixation and nonspecific motor responses are the only acquisitions in most [34, 59]. Poor weight gain, aspiration pneumonias and the need for gastrostomy are the rule [34]. Smith reports death usually before 2 years and one child surviving to 9 years, while the 2021 GeneReview states that few reach 10 years and the oldest known individual died at 17 years; the more recent figure is adopted [3, 34]. The Brazilian book also describes a cortex with four rather than six layers, grey matter heterotopias, spasticity, opisthotonus and a characteristic EEG [59].

3.5. Tubulinopathies

Motor and cognitive impairment is present in almost all and correlates with malformation severity: lissencephaly, microlissencephaly and severe generalised dysgyria are associated with spastic tetraplegia, almost no voluntary motor control and absent eye contact, whereas mild to moderate dysgyria is associated with mild motor disability and ID [5]. Most have severe to profound ID and few have near-normal development [5]. Epilepsy is highly variable and does not necessarily depend on severity, gene or variant; early epileptic encephalopathy with spasms is common in lissencephaly and severe generalised dysgyria, and seizures occur in fewer than 30% of patients with a simplified gyral pattern [5]. Facial diplegia and strabismus suggestive of pseudobulbar palsy are frequent in central pachygyria and various dysgyrias, and TUBB3 is associated with CFEOM type 3 with or without cortical malformation or axonal polyneuropathy [2, 5]. Mild forms survive into adulthood and severe forms may die early from seizures or pneumonia [5].

3.6. XLAG and other ARX forms

Individuals with XLAG are genotypically male and present with severe congenital or postnatal microcephaly, lissencephaly, corpus callosum agenesis, intractable neonatal epilepsy, poor temperature regulation, chronic diarrhoea and ambiguous or underdeveloped genitalia, with high lethality in the first three months [4, 16, 71]. Females heterozygous for ARX loss of function are asymptomatic or have delay, ID (55% of heterozygotes), learning difficulties and behavioural abnormalities, with seizures in 64% of those with ID; imaging frequently shows corpus callosum agenesis, without reported lissencephaly [3, 80].

3.7. Reelin pathway: RELN LCH and VLDLR cerebellar hypoplasia

RELN LCH combines classical-type lissencephaly with an anterior gradient, profound cerebellar hypoplasia without detectable folia and hippocampal and brainstem malformation, with hypotonia, ataxia, severe cognitive delay and seizures; complete reelin absence causes a consistently severe picture, while reduced expression may leave the cerebellum nearly normal [6, 15]. VLDLR cerebellar hypoplasia is congenital non-progressive ataxia, mainly axial, with delayed walking (frequently after 6 years) or never independent walking, moderate to profound ID, dysarthria, strabismus and seizures [6]. In more than 50 reported individuals, cerebellar hypoplasia was present in 44/44, simplified gyration in 43/44, pontine hypoplasia in 36/44, ataxia in 53/53, dysarthria in 34/42, strabismus in 40/51, hypotonia in 28/37, short stature in 19/42, microcephaly in 8/38 and epilepsy in 7/53 [6]. Some patients from Turkey displayed quadrupedal locomotion, an environment-dependent behavioural adaptation; most follow simple commands and adults do not live independently [6]. In Hutterites, seizures occurred in 40% and community experience suggests little reduction in life expectancy, without a formal study [6].

3.8. Epilepsy: aetiology-specific series

The clinical course is most strongly determined by epilepsy. In a retrospective German study involving 16 centres, all 22 patients with PAFAH1B1 lissencephaly had drug-resistant epilepsy, beginning within the first 6 months in 82%, most frequently with spasms, and subsequently evolving to focal and generalised tonic–clonic seizures [39]. In the cohort of 47 patients described in 2024, all but two had epilepsy, with median onset at 6 months (2.1 to 42 months) and epileptic spasms as the first manifestation in 70%; the same group's 2026 publication cites 57% for this proportion, and the figure from the 2024 primary source is adopted [33, 81]. Patients without spasms achieved significantly more developmental milestones [81]. In a Swedish cohort of 20 patients with epilepsy and the lissencephaly spectrum, PAFAH1B1 or 17p13.3 deletions explained 55%, tubulins (TUBA1A, TUBG1, TUBGCP6) 20%, and imaging reliably predicted the molecular result only in agyria with a posterior–anterior gradient, without consistent correlation with clinical severity or treatment response [82]. In female DCX SBH, LGS occurs in approximately half of patients with thick or thin bands [4]. Treatment responses in these series are discussed in section 7.

3.9. Lissencephaly 9 caused by MACF1

Among the eight children of the founding series (seven families of Indian, Dutch, North American, Polish, Hispanic, Syrian and Japanese origin; ages at last assessment 2.5 to 16 years), all had global delay, severe ID, axial hypotonia and seizures, and three had limb spasticity [8]. The monozygotic twin girls, non-verbal, had the most severe ID, short stature and microcephaly (−3 to −5 SD); one child spoke more than 40 words, another used an augmentative communication device and the rest had few or no words [8]. Seizures began between 3 and 7 months in seven of the eight (at 5 years in the Irahara patient), with infantile spasms in two (resolved with corticosteroid therapy in one), myoclonic seizures in two and focal and generalised seizures in the remaining four, evolving to Lennox–Gastaut syndrome in three [8]. Involuntary movements (chorea, athetosis, dystonia and, in the twins, ballismus) occurred in three and stereotypies in two; cranial nerve deficits consisted of dysphagia in four (two with gastrostomy), facial asymmetry in one and strabismus in three, and the Irahara patient had limited ocular abduction at 2.5 years, normalised by 5 years [8]. Two children had cortical visual impairment and one left optic nerve hypoplasia; hearing was normal in all [8]. One child had poor thermoregulation and neurogenic bladder and bowel attributed to spinal cord dysfunction without a structural lesion, and another a ventricular septal defect that closed spontaneously; there was no consistent facial dysmorphism and no other malformations [8].

In the 2025 review of the 18 individuals with variants in the EF1-EF2-GAR module (9 from the new cohort and 9 previously published), global delay was present in 18 of 18, hypotonia in 16 of 18 and severe ID in 12 of 12 assessed; 6 of 16 never sat (the others between 1 and 1.5 years), 8 of 17 never walked (the others between 1 year 8 months and 7 years) and 10 of 18 did not speak, with at most single words and only one individual able to form short sentences [9]. Epilepsy appeared in 16 of 18 (89%), in the first year in 15 of 18 (83%), with infantile spasms in 9 of 18 (50%), focal or generalised tonic–clonic seizures in 7 of 18 (39%) and Lennox–Gastaut syndrome in 4 of 18 (22%), a proportion possibly underestimated because of the young age of some patients [9]. HC was normal in 13 of 17, with microcephaly in 3 (−2.1 to −5 SD) and borderline macrocephaly in 1 [9]. The child homozygous for p.(Arg24Ter) had severe ID, epilepsy and postnatal microcephaly (HC 33 cm at birth, −1.27 SD, and 46 cm at 6.5 years, −4.48 SD) [9].

3.10. Lissencephaly 10 caused by CEP85L

The 2020 series brought together 13 patients from 9 families, mostly adults (19 to 52 years) and two children aged 3 and 5 years [11, 12]. Two had posterior agyria with an abrupt transition to mild anterior pachygyria, two pachygyria and nine SBH, always posterior-predominant in the 12 with reviewed MRI [12]. Intellectual development was normal in 4 of 13, with mild ID in 3, moderate in 5 and severe in 1 (IQ 58 to 84 in the four assessed, at least one attending university); 12 of 13 walked and all acquired language, with delayed first words (18 months to 4 years) in 6 [12]. Epilepsy appeared in 12 of 13 (92%), with onset between 15 months and 14 years (median 5 years) and drug resistance in all but three; some had normal early development with regression after the onset of refractory seizures [12]. OMIM reports seizure onset from the first year of life, whereas the article's table gives 15 months as the minimum age [11, 12]. Four patients had visual problems (cortical visual impairment, visual agnosia and myopia), and three adults from family V had psychiatric disorders (depression, bipolar disorder, hallucinations, aggressiveness and suicide attempts) [11].

In the parallel 2020 series of seven patients, bilateral posterior pachygyria had a thin cortex, an underlying cell-sparse zone and a thick subcortical band in the parietal, occipital and temporal lobes, completely sparing the cortex anterior to the central sulcus; all had reduced white matter and a dysmorphic corpus callosum, two a Chiari I malformation, and all walked [61]. Seizures were focal or epileptic spasms, four had strabismus, none had cortical visual impairment, HC was normal and there was no consistent facial dysmorphism [61]. A 3-year-old Italian child with c.232+1del had mild psychomotor delay, subtle dysmorphism (bitemporal narrowing, prominent ears and posterior plagiocephaly) and, on MRI at birth, temporo-parieto-occipital lissencephaly with periventricular SBH [78]. In a Canadian family, the p.(Thr66Ala) variant segregated with epilepsy across three generations; the proband, with epilepsy and delay at 5 years, had posterior pachygyria, and affected relatives had similar SBH [79].

The 2025 review of 32 individuals, 31 with postnatal assessment, recorded global delay in 71% (22/31), speech delay in 11 and motor delay in 6, ID in 74% (23/31), focal or generalised seizures in 90% (28/31) and associated SBH in 55% (17/31) [21]. It also described the first prenatal diagnosis and a mosaic variant, and observed that stronger splicing variants and constitutional variants tend to cause more severe presentations than mosaic variants [21]. In the Italian series, c.232+1 and start-loss variants were associated with the most severe imaging, and +3 to +5 variants, which leave some normal transcript, with mild presentations [78].

4. Clinical Recognition and Warning Signs

4.1. Prenatal period

The normal fetal brain contour is smooth until approximately 24 weeks, so lissencephaly cannot be reliably recognised on a second-trimester anomaly scan; the Oxford Desk Reference places ultrasonographic detection from 24 weeks and considers ultrasonographic diagnosis impossible before approximately 28 weeks [16, 83]. The PAFAH1B1 GeneReview states that level 2 ultrasonography does not detect lissencephaly, and the DCX GeneReview indicates that, without family history, DCX lissencephaly is recognisable only after 27 weeks, if at all [3, 4]. The European consensus adds that under- or overmigration malformations are difficult to recognise before 20 weeks, the legal termination limit in many countries, and any prenatal suspicion requires postnatal confirmation [1]. Prenatal warning signs are indirect: ventriculomegaly, delayed opercularisation, microcephaly, growth restriction, polyhydramnios and reduced fetal movements [2, 4, 33]. In the German cohort, 21 of 38 PAFAH1B1 patients had undergone routine ultrasound screening, suspicious in only 4 (19%), one at 22 weeks [33]. In a 2026 Chinese cohort, 17 of 20 fetuses with imaging-suggested lissencephaly had concomitant abnormalities, mainly ventriculomegaly (50%) and microcephaly (25%), and half of the 20 children diagnosed after birth had had no prenatal abnormalities; in the review of 154 literature cases, 48% had no specific prenatal findings [73]. Fetal MRI is more accurate than ultrasonography, and associated findings (congenital microcephaly, hydrocephalus, cerebellar hypoplasia) may guide earlier recognition; in MDS, prenatal ultrasonography and MRI may show lissencephaly, colpocephaly and corpus callosum dysgenesis [16, 65].

4.2. Neonate and infant

Lissencephaly should be suspected in a neonate or infant with hypotonia, poor feeding or weak sucking, poor head control, little visual interaction and, later, epileptic spasms; HC is normal or slightly reduced at birth and subsequently decreases [3, 16]. Lissencephaly may be seen on cranial ultrasonography requested for poor feeding, hypotonia or seizures, but MRI is the investigation of choice [16]. In the German cohort, the first symptom in PAFAH1B1 patients was a seizure in 46%, delay in 15% and hypotonia in 15%, and median age at diagnostic suspicion was 5 months [33]. Characteristic facial dysmorphism, particularly a high forehead with vertical furrows when crying and bitemporal narrowing, in a hypotonic infant should suggest MDS [34]. A boy with neonatal encephalopathy, intractable seizures, hypothermia, diarrhoea and ambiguous or underdeveloped genitalia should raise suspicion of ARX XLAG [16, 84]. XLAG and MDS are among genetic causes to consider in neonatal encephalopathy with intractable seizures [84]. Before counselling the family of a child with infantile spasms, the Oxford Desk Reference recommends excluding structural malformations such as lissencephaly and periventricular nodular heterotopia [85].

4.3. Child, adolescent and adult

In girls and women, DCX SBH should be considered with drug-resistant focal or generalised epilepsy, LGS, ID and behavioural disorders, but also with mild learning difficulties and late-onset seizures [4, 86]. In X-linked intellectual disability, the Oxford Desk Reference recommends asking about seizures in males and females in the family, because SBH in DCX carriers and nodular heterotopia in FLNA carriers provide gene clues [86, 87]. A family history compatible with X-linked transmission (affected males connected through asymptomatic or less affected females, without male-to-male transmission), with epilepsy, miscarriages, stillbirths or children dying early without diagnosis, reinforces suspicion, but its absence does not exclude DCX [4]. In boys with late-onset focal seizures and MRI showing thin SBH or frontal pachygyria, DCX mosaicism is a possibility to investigate [4]. A child with congenital non-progressive ataxia, very delayed walking and strabismus, particularly of Hutterite ancestry, should suggest VLDLR cerebellar hypoplasia, a differential of ataxic cerebral palsy [6, 88].

4.4. Warning and emergency signs

Suspected infantile spasms in an infant younger than 2 years are a neurological emergency: National Institute for Health and Care Excellence (NICE, United Kingdom) guideline NG217 recommends seeking advice and urgent referral to tertiary paediatric neurology within 24 hours for sleep EEG and rapid treatment, and the German-speaking Society of Neuropediatrics (GNP) guideline recommends EEG within the first days and treatment immediately after confirmation, including when aetiology is already known [37, 38]. Other warning signs are developmental regression accompanying worsening epilepsy or nutrition; choking, chronic cough, prolonged meals (more than 30 minutes), poor weight gain and recurrent respiratory infections, indicating aspiration; status epilepticus and pneumonia, the main causes of death; and new unusual episodes, warranting neurological consultation and EEG [3, 4, 33]. In an infant with lissencephaly treated with vigabatrin, new hypotonia, somnolence or respiratory failure should prompt urgent MRI, because co-occurrence of vigabatrin-associated MRI abnormalities with acute obstructive hydrocephalus requiring emergency external ventricular drainage has been described [89]. In XLAG, hypothermia and diarrhoea with dehydration require intensive support [15].

4.5. Screening

There is no newborn screening for lissencephalies or MDS. In Hutterite communities, the carrier frequency of the VLDLR deletion, estimated at 1 in 15, may justify population reproductive screening through targeted founder deletion analysis [6, 55].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Imaging criteria and definitions

There are no formal clinical criteria; diagnosis rests on imaging (or neuropathological) demonstration of the malformation and, whenever possible, identification of the molecular cause [1, 2]. The Neuro-MIG consensus defines the lissencephaly spectrum using Human Phenotype Ontology (HPO) terms: agyria (HP:0031882) and pachygyria (HP:0001302) correspond to a gyral pattern with absent or broad gyri associated with an abnormally thick cortex, and SBH (HP:0032409) to a grey matter band separated from the cortex and ventricles by white matter [2]. Normal cortical thickness ranges from 1 to 4.5 mm in pathology, and primary and secondary sulci are normally less than 1.5 cm apart; MRI measurements should not be used as rigid thresholds [1]. In “thick” lissencephaly, the cortex measures 10–20 mm and has four layers; in “thin” lissencephaly it measures 5–10 mm, still above normal; “variable” lissencephaly mixes thick and thin areas [1, 58]. In agyria, the brain has a “figure-of-eight” shape with broad vertical Sylvian fissures; at all severities the grey–white matter junction is smooth, white matter volume reduced, commissures thin and ventricles enlarged [1]. Thick SBH measures more than 7 mm and lies in deep white matter, while thin SBH measures 1 to 7 mm and is subcortical [1, 58]. Sources are not uniform: Di Donato 2017 Table II and the 2020 consensus DCX pattern use 5 mm as the boundary, and the DCX GeneReview describes thick bands as >8 mm and thin bands as 4–7 mm in one section and 1–7 mm in another, so each value should be read with its source definition [2, 4, 58].

Di Donato's algorithm first separates SBH (relatively preserved six-layer architecture, strongly associated with mosaicism) from lissencephaly, classifying mixed frontal pachygyria with posterior bands in the lissencephaly group; it then defines the gradient by gyral width (diffuse, posterior > anterior including perisylvian, anterior > posterior, or temporal > posterior > anterior, only in boys with ARX), cortical thickness and finally non-cortical malformations, of which cerebellar hypoplasia is most frequent and easiest to assess [58]. Cortical thickness can be measured reliably only before 3 months (heavily T2-weighted, ideally 3D) or after 24 months (3D T1), because ongoing myelination makes measurements imprecise between these ages [58].

5.2. Imaging protocol

CT is discouraged for radiation protection and reserved as second-line for calcifications; MRI is the most important investigation, preferably at 3 T with optimised protocols [1]. In the myelinated brain, the protocol includes 3D T1 with slices of 1 mm or less, axial and coronal T2 at 3 mm or less and FLAIR (fluid-attenuated inversion recovery), preferably 3D; in neonates, T2 in all three planes is essential and FLAIR unnecessary; diffusion tensor imaging is useful [1]. Because the “T2-isointense” phase between 8 and 12 months may conceal malformations, the consensus recommends repeating MRI after myelination is complete (in the 3rd year of life) if the first was performed in that phase; the diagnostic consensus also recommends repeating MRI performed between 3 months and 2.5 years or of low quality when diagnosis remains unresolved [1, 2]. MRI requests should include HC, syndromic signs, seizure semiology and EEG [1]. In Regensburg, specialist MRI review before genetic testing was associated with a yield of 37% (117 patients) versus 18% without review (784 patients), unpublished internal data cited by the consensus [2].

5.3. Imaging pattern and probable gene

The 2017 classification, built on 188 patients from the Seattle (129, North America) and Florence (59, Europe) cohorts, excluded cobblestone malformation; 78% had thick classical lissencephaly, and the most frequent pattern was posterior-predominant partial agyria-pachygyria (29.2%), almost exclusively caused by severe PAFAH1B1 variants [58]. The following table condenses the pattern–gene correspondence from Di Donato 2017 Table II and Di Donato 2018 Table 1, with observed yield in the series of 811 patients [7, 58].

The 2020 consensus highlights four patterns with yield above 90% in one or two genes: diffuse agyria with cortex >10 mm (PAFAH1B1, mainly 17p13.3 microdeletion, and DCX); occipital agyria with frontal pachygyria (PAFAH1B1, rarely TUBG1 and TUBA1A at codon 402); temporal pachygyria with a 5–10 mm cortex, complete corpus callosum agenesis and severe hypomyelination (ARX; DYNC1H1 produces a similar pattern without hypomyelination); and diffuse SBH with band >5 mm, pathognomonic of DCX in females and males [2]. Tubulinopathies are recognised by basal ganglia dysmorphism (enlarged caudate, absent or diminutive anterior limb of the internal capsule, “hooked” frontal horns), callosal abnormalities, vermian dysplasia with “diagonal” folia obliquely crossing the midline, and asymmetrical brainstem hypoplasia with a midline ventral indentation [2, 5]. In DCX lissencephaly, the cortex has an anterior gradient, the caudate head may be enlarged and the cerebellum and brainstem are normal; in female SBH, bands are symmetrical, usually bilateral and isointense to grey matter [4]. In PAFAH1B1 ILS, ventricles are enlarged mainly posteriorly, the corpus callosum is mildly hypoplastic and cavum septi pellucidi et vergae may occur; PAFAH1B1 SBH is confined to parietal and occipital lobes, while diffuse or frontal SBH points to DCX [3]. Imaging alone does not reliably distinguish genetic from non-genetic causes in MCD generally [2]. Lissencephaly differs: the malformation is considered exclusively genetic, and for non-specialists recognising lissencephaly suffices to initiate investigation [2, 7].

The LIS9 pattern does not appear in the 2017 and 2018 tables and requires targeted reading of the brainstem. In the cortex, the 2018 series recorded diffuse pachygyria in all, with a posterior > anterior gradient in six of eight and no gradient in the twins, mildly thickened cortex in six (4–7 mm in the article's table; 7–10 mm in the text) and thick cortex (10–15 mm) in the two patients with p.Asp7186Tyr [8]. All had dysplastic hippocampi, a thin anterior commissure and reduced white matter; the corpus callosum was thin in five of eight in the article's table (six in the text), and the hippocampal commissure and optic chiasm were normal [8]. On the midsagittal section, the midbrain is mildly narrow, the pons very narrow, with small or no ventral bulges, and the medulla mildly thick; the basis pontis measured 4–5 mm in three patients, 2–3 mm in three and was absent in two [8]. On axial sections, the lower midbrain, pons and medulla are narrow in the dorsoventral axis, with a deep median ventral cleft, and very broad in the transverse axis (the medulla almost twice the usual width), a configuration resembling a thick “W”; the cranial nerves emerge from the medial and lateral ventral surfaces of the dysplastic brainstem [8]. 3D coronal reconstructions along the ventral surface of the brainstem showed absent or very scarce transverse pontine fibres, leaving the pyramidal tracts exposed, and diffusion tensor imaging (DTI) in one patient showed small superior cerebellar peduncles and absence of the anterior and posterior transverse pontine fibres [8]. The vermis was moderately small in two, mildly small in five and normal in one, with foliar dysplasia in all [8]. The residual 3–4 mm ventral nodules, representing crossing pontine fibres, were larger with Cys7188 substitutions and with the deletion; anterior commissure hypoplasia was most severe with p.Cys7135Phe and lissencephaly most severe with p.Asp7186Tyr [8].

The 2025 cohort confirmed the pattern and added variations: residue p.Cys7130 gave diffuse pachygyria of intermediate severity with no gradient or with subtle anterior predominance, p.Cys7185Arg was associated with global cerebral atrophy, the β2 strand variant (p.Arg7147Pro) gave less severe pontine hypoplasia, a thin medulla without the “W” shape and dysgyria without lissencephaly, and the EF-hand variants reproduced the classical pattern, with more severe hypoplasia of the corpus callosum and vermis in one patient assessed at 52 years; the homozygous exon 1 variant gave anterior-predominant diffuse pachygyria, mild pontine hypoplasia, a less characteristic “W” shape and a normal cerebellum [9]. In practice, MRI of a lissencephaly should include thin axial sections of the brainstem and, if possible, DTI and coronal reconstructions of the ventral surface of the pons, because brainstem shape and the scarcity of transverse pontine fibres are the features that distinguish LIS9 [8, 9].

5.4. Genetic investigation algorithm

The Neuro-MIG consensus recommends a concerted effort towards aetiological diagnosis in all MCD, because diagnosis ends the diagnostic odyssey, informs prognosis and recurrence risk, predicts epilepsy surgery outcome and guides management; final interpretation falls to the clinical geneticist, ideally after discussion with a molecular geneticist, neuroradiologist and neurologist [2]. The flow chart includes prenatal and perinatal history, pedigree with consanguinity, specialist MRI review, exclusion of congenital cytomegalovirus infection when suggested, CMA and NGS, segregation and counselling [2].

In practice before arrays, MDS with normal high-resolution karyotype was confirmed by FISH with 17p13.3 critical region probes, and the Brazilian book describes diagnosis by karyotype, array comparative genomic hybridisation (array-CGH) or FISH [34, 59]. For lissencephaly with suggestive MRI, GeneReviews proposes PAFAH1B1 sequencing followed by deletion/duplication analysis or a multigene panel and, when the phenotype is indistinguishable from other causes, exome or genome sequencing [3]. For DCX, in a male with characteristic MRI and X-linked history, DCX sequencing first; genomic testing is preferred without characteristic imaging or family history; if several females in the family have SBH and no germline variant is found, mosaicism should be investigated [4]. In a boy with apparent ILS and normal DCX and PAFAH1B1 tests, maternal MRI may reveal clinically inapparent SBH, but because carriers may have normal MRI, DCX testing should precede maternal MRI [16]. The 2026 Chinese cohort recommends exome sequencing as first-line testing [73]. In a Uruguayan case, a pregnancy with exposure to multiple drugs had directed diagnosis towards an acquired cause, and only genome sequencing identified the PAFAH1B1 variant; suggestive imaging would have enabled targeted testing [90].

5.5. Test yield and what the percentages mean

GeneReviews Table 1 indicates the proportion of pathogenic variants detectable by each method, which is not clinical test sensitivity: in PAFAH1B1, sequencing detects 77% and deletion/duplication analysis 23%, with balanced translocations rare; in DCX, sequencing detects 94–96% and deletion/duplication analysis 4–6%; in VLDLR, 85% and 15% [3, 4, 6]. Clinical sensitivity also depends on pattern: in partial posterior SBH, only 38% of patients were solved, and in the panel series only 2 of 15 posterior SBH cases had a variant, both mosaic PAFAH1B1 [7]. Even in thick diffuse SBH, at least 20% of patients test negative for DCX [58].

5.6. Test limitations

A negative result does not exclude a genetic cause, for identifiable technical reasons. Targeted deletion/duplication analysis (quantitative PCR, long-range PCR and multiplex ligation-dependent probe amplification, MLPA) may fail to define the breakpoints of large deletions or detect deletion of adjacent genes such as YWHAE, and therefore does not distinguish ILS from MDS; CMA is the appropriate test for this purpose [3]. Balanced reciprocal translocations disrupting PAFAH1B1 or DCX are not detected by CMA or by sequencing [3, 4]. DCX sequencing does not detect exonic or whole-gene deletions in females, although in males a deletion may cause amplification failure [4]. Exome and genome sequencing detect copy number by read depth with possibly lower sensitivity than targeted methods, and the report should state whether exon-level copy number variant (CNV) analysis was performed [2, 4]. Standard-coverage exome sequencing on blood misses low-level postzygotic mosaicism, described in DCX, PAFAH1B1, FLNA, TUBB2B and PI3K-AKT-mTOR pathway genes, and the usual approach is suitable only for germline variants and mosaics above approximately 30% of cells [2]. Tubulin genes are prone to misalignment: the TUBB2B variant NM_178012.5:c.743C>T p.(Ala248Val), described as pathogenic and with conflicting classifications in ClinVar [63], appears in the gnomAD population database (Genome Aggregation Database) at a frequency of 3% flagged as an artefact and may be filtered out, so the consensus suggests Sanger sequencing of TUBB2B in suggestive unresolved cases [2]. RELN testing must include copy number analysis, because single-exon deletions have been found [7]. For ARX, the assay must detect small in-frame duplications in a GC-rich repeat [15]. Data on deep intronic DCX variants are insufficient [4]. The consensus recommends classifying as benign a variant with a frequency above 1% in gnomAD (more restrictive than the general 5% of the American College of Medical Genetics and Genomics and Association for Molecular Pathology, ACMG/AMP, criteria) and considering the presence of five or more homozygotes as suggestive of benignity, checking for hemizygosity due to deletion [2].

In MACF1, the in-frame intragenic deletion of exons 62 to 93 in the Irahara patient was not detected by exome sequence variant analysis and appeared only on manual inspection of read depth and with CNV-calling algorithms on the exome (GATK4) and structural variant callers on the genome (Manta), with breakpoints confirmed by PCR; with a typical phenotype and a negative exome, targeted review of coverage of the C-terminal exons is therefore justified [8]. The existence of transcripts with distinct exons 1 means that a variant in a brain-specific exon 1 (such as NM_012090.5 c.70C>T) is not annotated on the MANE transcript, and an analysis restricted to that transcript may miss it [9]. Interpretation of variants outside the EF1-EF2-GAR module requires caution, because MACF1 is a large gene with expected neutral missense variation, including occasional de novo and biallelic variants; the 2025 authors recommend not considering a MACF1 variant causal without several observations in the same domain or functional studies [9].

In CEP85L, start-loss variants affect only NM_001042475, and the 2025 review recommends this transcript for interpretation, because annotation on the alternative transcript causes confusion [12, 21]. Mosaic variants, already described, may be missed at low read depth [21]. CEP85L explained 3 of 14 (21%) posterior lissencephalies unresolved after study of the known genes and 3 of 36 unresolved lissencephalies, pachygyrias or SBH in a research exome, and should be included in lissencephaly panels [12, 61, 78].

5.7. Report content and interpretation

Molecular diagnosis is established by an ACMG/AMP pathogenic or likely pathogenic variant, heterozygous in PAFAH1B1, hemizygous in a male or usually heterozygous in a female in DCX, or biallelic in recessive genes; a variant of uncertain significance (VUS) neither establishes nor excludes diagnosis [3, 4, 6]. The report should include the gene, versioned transcript (preferably MANE Select), Human Genome Variation Society (HGVS) DNA and protein description, zygosity or mosaic allele fraction, ACMG/AMP class, all pathogenic variants and all VUS in MCD genes, indication of CNV analysis and recommendation for segregation testing; the consensus encourages sharing VUS in ClinVar or the Leiden Open Variation Database (LOVD) [2]. Gene lists should be updated every 6 months [2]. In the series of 811 patients, 12 ACMG/AMP VUS were considered causal by the authors because of imaging concordance and absent parental samples, illustrating the weight of imaging phenotype in reclassification [7]. For highly specific patterns with negative testing, the consensus recommends offering participation in gene-discovery research [2].

5.8. Other investigations

EEG in classical lissencephaly may show high-amplitude diffuse fast rhythms, considered highly specific; the Brazilian book reports them in 80 to 99% of cases, without a denominator [3, 59]. Prolonged video-EEG may be required to characterise epileptic burden or uncertain events [4]. In the differential diagnosis, the Oxford Desk Reference recommends ophthalmological examination (retinal dysplasia, cataract and myopia in dystroglycanopathies), plasma creatine kinase (CK) (Walker–Warburg, muscle–eye–brain disease, Fukuyama congenital muscular dystrophy), very-long-chain fatty acids (peroxisomal disorders with lissencephaly or polymicrogyria) and nerve conduction studies if hyporeflexia or foot deformities are present (DYNC1H1) [16]. The consensus adds metabolic investigations in microcephaly, polymicrogyria and cobblestone malformation, and muscle biopsy with weakness or elevated CK [2]. After epilepsy surgery, autopsy or sudden unexpected death in epilepsy, neuropathological examination, biobanking and genetic testing are recommended, following the 2016 International League Against Epilepsy (ILAE) neuropathology task force procedure [2].

5.9. Prenatal diagnosis and preimplantation testing

When the familial variant or parental rearrangement is known, prenatal diagnosis by chorionic villus sampling at 11 weeks or amniocentesis and preimplantation genetic testing (PGT) are possible [3, 4, 16]. Without a known variant, fetal imaging has the limitations described in section 4, and the 2026 Chinese cohort proposes an integrated prenatal imaging and fetal exome model, with a yield of 55% and a warning that lissencephaly signs appear late in gestation [73]. In a female fetus carrying a DCX variant, manifestations cannot be predicted, so prenatal testing should be offered only after detailed discussion of reduced penetrance and variability [4]. SBH is recognised only after birth in most cases, although occasionally detected by late fetal MRI or ultrasonography [4].

6. Differential Diagnosis

6.1. Recognising the malformation

The greatest diagnostic difficulty is correctly recognising the malformation. Severe congenital microcephaly with reduced gyration, cobblestone malformation and polymicrogyria, including polymicrogyria-like tubulinopathy variants, are mistaken for lissencephaly, leading to inefficient testing and incorrect counselling [3]. Clinically, lissencephaly is associated with normal or slightly reduced birth HC (above −3 SD) and diffuse hypotonia, with slightly increased tone only at wrists and ankles; severe primary microcephaly with gyral abnormalities has birth HC of −3 SD or less, with hypotonia or spasticity, and polymicrogyria, particularly frontal, is associated with spastic tetraparesis [3]. The Brazilian book emphasises that patients with MDS are born with normal HC and remain hypotonic, unlike other conditions with prenatal microcephaly and spasticity [59]. MRI or autopsy neuropathology is necessary for confirmation [3].

6.2. Comparative table

6.3. Miller–Dieker syndrome versus isolated lissencephaly and neighbouring deletions

Clinically, MDS differs from ILS through significant facial dysmorphism and a more severe lissencephaly grade, and molecularly through deletion extension beyond the 400 kb critical region, including YWHAE and frequently CRK [54]. In practice, deletion of PAFAH1B1 alone or with only proximal genes causes ILS, while deletion extending distally to YWHAE causes MDS; in the series of 811 patients, the 75 PAFAH1B1-YWHAE deletions clustered in the diffuse agyria pattern [7]. 17p13.3 deletions including YWHAE but sparing PAFAH1B1 constitute an entity distinct from MDS: in a 2023 series of 4 new patients and 19 from the literature, phenotype comprised facial dysmorphism (prominent forehead, epicanthal folds, broad nasal root), developmental delay, leukoencephalopathy with cystic white matter abnormalities (probably dilated perivascular spaces), frequent epilepsy and ID, without lissencephaly [66]. The authors also cite six patients with 17p13.3 microdeletion sparing YWHAE, CRK and PAFAH1B1, with white matter abnormalities and dysmorphism without ID or seizures [66]. GeneGuidance curation cites a 2023 series of ten individuals with heterozygous YWHAE loss of function (three point variants and seven deletions below 1 Mb sparing PAFAH1B1), with delay, speech delay, seizures and malformations such as corpus callosum hypoplasia and delayed myelination, milder than with larger deletions [15, 92]. Orphanet records distal 17p13.3 microdeletion with 16 validated cases [31]. 17p13.3 microduplications divide into class I, with YWHAE without PAFAH1B1, associated with upslanting palpebral fissures, synophrys, a square nasal tip, large ears, large hands and more frequent autism, and class II, with PAFAH1B1, associated with a broad midface, low-set ears, hypertelorism, mild to severe delay, hypotonia, growth restriction and corpus callosum hypoplasia; the GeneReview writes “PAFAH1A1” in the class I definition, a typographical error for PAFAH1B1 [3, 34]. Orphanet records 50 validated 17p13.3 microduplication cases [32].

6.4. PAFAH1B1 versus DCX versus tubulins

The gradient of the malformation is the most useful discriminator: almost all posterior lissencephaly results from PAFAH1B1 or tubulins (including TUBG1), and almost all anterior classical lissencephaly results from DCX or the actins ACTB and ACTG1; only DYNC1H1 and KIF5C produce both anterior and posterior forms [58]. SBH points to DCX, much more frequently than to PAFAH1B1, and SBH restricted to the posterior lobes points to mosaic PAFAH1B1 [3, 58]. A posterior lissencephaly or SBH with an abrupt transition anterior to the central sulcus, normal HC, sometimes normal cognition and vertical transmission points to CEP85L; in PAFAH1B1 lissencephaly, the anteroposterior transition is less abrupt and the presentation is severe, with no reproduction of affected individuals [12, 61]. In severe forms, the gradient may be difficult to discern, and near-complete agyria in a boy may result from an early truncating DCX variant [3, 7]. The presence of non-cortical malformations (dysmorphic basal ganglia, partial agenesis of the corpus callosum, enlarged tectum, brainstem and cerebellar hypoplasia) suggests tubulins or motor proteins, except for TUBA1A variants at codon 402 [58]. In Baraitser–Winter syndrome, epilepsy occurs in 53% and heart disease in 33%, and some patients have normal intelligence; short bands between frontal pachygyria and normal posterior cortex are characteristic [58, 91].

6.5. Lissencephaly with corpus callosum agenesis, ARX and its allelic disorders

Lissencephaly with corpus callosum agenesis in a 46,XY individual is typically caused by ARX, and differs considerably from PAFAH1B1 or DCX forms through microcephaly, neonatal-onset intractable epilepsy, poor temperature regulation, chronic diarrhoea and abnormal genitalia, sometimes already recognisable in a boy with delay and seizures [3, 71, 93]. Corpus callosum agenesis with lissencephaly also occurs in tubulinopathies (TUBA1A, TUBG1, TUBB3) and with KIF5C or KIF2A [16]. Allelic ARX disorders without cortical malformation, namely X-linked infantile spasms, DEE1 and Partington syndrome, enter the differential of any boy with spasms and normal MRI, and ARX is among the most frequent de novo causes of early-onset epileptic encephalopathy with an Ohtahara pattern, alongside STXBP1, KCNQ2 and SCN2A [15, 85].

6.6. Cobblestone malformation, microlissencephaly and others

Walker–Warburg syndrome, muscle-eye-brain disease and Fukuyama congenital muscular dystrophy are recessive dystroglycanopathies due to defective O-mannosylation; the first, the most severe, combines hydrocephalus, agyria, retinal dystrophy and sometimes encephalocele, with markedly raised CK [16, 94, 95]. Muscle-eye-brain disease is relatively common in Finland and Fukuyama dystrophy in Japan, where a 3 kb insertion in FKTN acts as a founder effect [16]. In a 2012 fetal series cited by the Oxford Desk Reference, 58 of 90 fetuses with cobblestone malformation (64%) obtained a molecular diagnosis, mainly in POMT1 (42%), POMT2 (17%) and POMGNT1 (17%) [16]. Microlissencephaly is associated with recessive genes such as NDE1 and WDR62, and Barth-type microlissencephaly has diffuse 4–8 mm agyria with severe brainstem and cerebellar hypoplasia, without an identified gene in the sources consulted [16, 58]. Pontocerebellar hypoplasias are distinguished from LCH by more severe vermian and pontine hypoplasia and by motor neuron degeneration or dyskinesia [6]. FLNA periventricular nodular heterotopia is lethal in many male fetuses, and approximately 10% of males with nodular heterotopia have mild FLNA variants; ARFGEF2 causes heterotopia with recessive microcephaly [16]. In the approach to inborn errors of metabolism, gyration abnormalities feature in Zellweger spectrum peroxisomal disorders, amino acid synthesis defects, O-CDG, congenital neuronal ceroid lipofuscinosis, glutaric aciduria type II, respiratory chain disorders and several disorders of cellular trafficking, especially Golgipathies; lissencephaly proper is associated with tubulinopathies, microtubule-associated proteins and actin genes (ACTB), and severe pachygyria with glutathione peroxidase 4 (GPX4) deficiency [96].

6.7. Lissencephaly with brainstem malformation: MACF1 and its differentials

In a lissencephaly with brainstem hypoplasia, brainstem shape guides diagnosis. In tubulinopathies, brainstem hypoplasia is asymmetrical, with ventral indentation, and is accompanied by dysmorphic basal ganglia, callosal anomalies and vermian dysplasia with “diagonal” folia; in LIS9, the pons and medulla are symmetrically narrow and broad, with a median ventral cleft and absent transverse pontine fibres [2, 5, 8]. In LCH caused by RELN or VLDLR, lissencephaly is thin and anterior-predominant and cerebellar hypoplasia dominates, whereas in LIS9 the predominance is posterior and the cerebellum is only mildly to moderately reduced [7, 8]. Posterior PAFAH1B1 lissencephaly and posterior-predominant CEP85L lissencephaly lack the brainstem malformation and the crossing defects described in LIS9 [8, 12]. The superior cerebellar peduncles in LIS9, sometimes prominent, are smaller than in the molar tooth sign of Joubert syndrome [8].

Outside the lissencephalies, the “W” configuration is more severe than the “butterfly” brainstem of horizontal gaze palsy with progressive scoliosis caused by ROBO3 and of split brain syndrome caused by DCC: all three genes produce a median ventral cleft, but only in LIS9 are the pons and medulla simultaneously very narrow and very broad; DCC variants are also associated with severe anterior commissure hypoplasia and variable agenesis of the corpus callosum, and biallelic ARHGEF2 variants with severe hypoplasia of the transverse pontine fibres with an enlarged medulla [8, 9]. These malformations, pontine tegmental cap dysplasia and Irahara syndrome itself are part of the differential diagnosis. The differential diagnosis of the biallelic anterior-predominant MACF1 form includes DCX in males and the actins ACTB and ACTG1, which do not produce the “W”-shaped brainstem [9, 58].

7. Treatments and Risk Management Strategies

7.1. Principles and recommendation sources

No disease-modifying treatment exists for any lissencephaly or MDS, and there are no specific clinical practice guidelines: the PAFAH1B1 and DCX GeneReviews explicitly state that there are no published consensus recommendations and their proposals are based on author experience [3, 4]. The 2020 Neuro-MIG consensus covers diagnosis and classification, not therapy, and observes that MCD medical management is rarely evidence-based [2]. The 2021 review by Koenig, Dobyns and Di Donato includes clinical follow-up recommendations and epilepsy management data, but only its abstract and highlights were accessible; these emphasise that intractable epilepsy is the principal clinical challenge and only symptomatic therapy exists to date [36]. Epilepsy recommendations therefore come from general IESS and LGS guidelines (GNP 2022 and NICE NG217, updated August 2026) and observational lissencephaly series; the 2012 American Academy of Neurology and Child Neurology Society (AAN/CNS) infantile spasm guideline is marked retired and cited only historically [37, 38, 97]. Management is symptomatic and multidisciplinary, coordinated by a paediatric neurologist experienced in multiple disabilities and refractory epilepsy [5, 16, 59]. Parents cope better when given accurate prognostic information early, and in severe lissencephaly discussing care limits, such as do-not-resuscitate orders, is usually appropriate in serious illness [3, 5].

7.2. Assessment after diagnosis

7.3. Infantile epileptic spasms: guideline recommendations

Epileptic spasms are sudden flexions or extensions, mainly proximal and axial, more sustained than myoclonus and less than a tonic seizure, occurring in clusters on awakening and potentially subtle; hypsarrhythmia is not mandatory for IESS diagnosis, the ILAE designation encompassing West syndrome [37, 98]. In the 2022 ILAE classification, lissencephaly is a structural-genetic aetiology of a developmental and epileptic encephalopathy, and each patient should be classified by syndrome and aetiology [98]. The 2022 German guideline, the third version of a GNP-coordinated document graded using SIGN (Scottish Intercollegiate Guidelines Network), applies to new-onset IESS, including when aetiology is already known [37]. Goals are rapid spasm cessation and resolution of hypsarrhythmia in wakefulness and sleep, assessed clinically and by EEG approximately 14 days after treatment initiation; clinical success is defined by 48 hours without spasms, and reduction alone does not count as response [37].

Guidelines diverge on first line: NICE offers prednisolone plus vigabatrin based on ICISS (International Collaborative Infantile Spasms Study), while GNP accepts hormonal treatment alone or combination treatment, because in ICISS the combination was better for spasm cessation at 2 weeks but did not show superiority in development or epilepsy at 18 months, and because combination adverse effects have been reported (movement disorders, encephalopathy, dysautonomia), particularly in trisomy 21 [37, 38]. Both documents reserve vigabatrin alone for tuberous sclerosis and corticosteroid contraindications [37, 38]. GNP relies on two level 1+ studies showing greater hormonal efficacy than vigabatrin, a conclusion already reflected in the 2012 US guideline [37, 97]. In diffuse lissencephaly there is no resectable focal lesion, limiting applicability of the GNP surgical recommendation, which requires focal signs on MRI, EEG, positron emission tomography (PET) or neurological examination [37]. NICE classifies combination vigabatrin and levetiracetam, nitrazepam and topiramate in IESS as off-label use in January 2025 [38].

7.4. Spasm drug dosage

Adverse effects and their management determine selection. Vigabatrin has a US boxed warning for permanent vision loss: bilateral concentric visual field constriction, possible central retinal damage, risk increasing with dose and cumulative exposure without a risk-free dose; baseline visual assessment (within 4 weeks after starting), at least every 3 months and 3 to 6 months after discontinuation is required, and availability is restricted to a risk evaluation and mitigation strategy (REMS) [100]. The European Kigabeq® summary of product characteristics (SmPC) describes irreversible visual field defects in approximately one-third of patients and recommends baseline assessment and at least every 6 months, and 6 to 12 months after discontinuation; perimetry is possible only above a developmental age of 9 years, so electroretinography is used in infants [102]. GNP notes that children with IESS have visual risk even without vigabatrin and electroretinography requires sedation [37]. In a retrospective study of 205 infants with spasms, 22% treated with vigabatrin and 4% treated with other therapies had symmetrical increased T2 signal and restricted diffusion in the thalamus, basal ganglia, brainstem and cerebellum, generally reversible on discontinuation [100, 102]. In a 4-month-old infant with PAFAH1B1 lissencephaly treated with vigabatrin, these abnormalities coexisted with acute obstructive hydrocephalus requiring emergency external ventricular drainage; the authors discuss a possible influence of the malformation [89]. With ACTH or glucocorticoids, GNP recommends: avoid live vaccines from 4 weeks beforehand (minimum 2) until 8 weeks afterwards, and inactivated vaccines during the preceding and following weeks; assess blood pressure, full blood count, glucose, electrolytes, liver and kidney function and echocardiogram before treatment; consider gastric protection; early antibiotics for infection; avoid varicella exposure and, after exposure, give passive immunisation within 72 hours without awaiting serology; measure blood pressure weekly and, from the 4th week, twice weekly, and glucose weekly [37]. Hypothalamic–pituitary–adrenal axis suppression may last up to one year, warranting a steroid card during that period and stimulation tests after prolonged treatment [37]. NICE also recommends varicella-zoster antibody testing before prednisolone and weekly blood pressure and glycosuria monitoring [38].

7.5. Epilepsy after the first year and chronic epilepsy

The most cited recommendation for chronic lissencephaly epilepsy, valproate and lamotrigine polytherapy, rests on a single retrospective German multicentre study (Herbst 2016), without a control group or randomisation [15, 39]. In that study of 22 patients with PAFAH1B1 lissencephaly, lamotrigine and valproate were rated most effective, with good or partial response in 88–100%, significantly better than levetiracetam, and the authors concluded patients may benefit most from lamotrigine, valproate, vigabatrin or phenobarbital [39]. The PAFAH1B1 GeneReview adds that even so, two-thirds continue having daily seizures [3]. In the cohort of 47 patients published in 2024, standard spasm protocols (hormonal treatment and/or vigabatrin) were most effective, with seizure control in 47%; for subsequent seizures, valproate and lamotrigine were most effective, followed by vigabatrin and phenobarbital, with seizure freedom in 20% [81]. In the Swedish cohort of 20 patients, the best responses were obtained with ketogenic diet, vigabatrin, clobazam, phenobarbital and valproate [82]. In the DCX GeneReview, data are even more limited: in two cases, spasms were controlled with phosphocreatine, prednisone, topiramate and nitrazepam, with drug-resistant recurrence at 3 years; one child from a lissencephaly cohort received ACTH, vigabatrin, levetiracetam, phenobarbital and topiramate without efficacy data; and, in a cohort of 47 patients with lissencephaly, valproate and lamotrigine were most effective according to unpublished author data [4]. The same GeneReview recommends balancing seizure control with adverse effects and limiting drug numbers, and both GeneReviews recommend prompt aggressive seizure treatment, because poor control worsens feeding, increases gastrostomy and pneumonia likelihood and compromises development [3, 4]. A 2025 review, citing a retrospective MDS study, reports perampanel efficacy in only 50% of patients [64].

When epilepsy evolves to LGS, NICE recommends involvement of an experienced neurologist and proposes sodium valproate first-line, with UK Medicines and Healthcare products Regulatory Agency (MHRA) safety measures; lamotrigine second-line, off-label in the United Kingdom as monotherapy below 13 years and adjunct below 2 years; third-line adjunctive cannabidiol with clobazam (≥2 years), clobazam, fenfluramine (≥2 years), rufinamide or topiramate (with a pregnancy prevention programme); then ketogenic diet and finally felbamate, unlicensed in the United Kingdom [38]. The same guideline lists carbamazepine, gabapentin, lacosamide, lamotrigine, oxcarbazepine, phenobarbital, pregabalin, tiagabine and vigabatrin as drugs that may worsen LGS seizures [38]. A divergence therefore exists between observational lissencephaly series reporting good response to lamotrigine, phenobarbital and vigabatrin, and the general LGS guideline identifying them as potential aggravators; decisions should be individualised and reassessed according to seizure evolution [38, 39, 81].

Valproate plus lamotrigine requires attention to US label boxed warnings. Valproate (divalproex, Depakote®) has a warning for fatal hepatotoxicity, generally within the first 6 months, with considerably increased risk in children younger than 2 years, particularly with multiple anticonvulsants, congenital metabolic diseases, severe epilepsy with ID and organic brain disease, a profile including infants with lissencephaly; in this group it should be used with extreme caution and as monotherapy, with liver tests before and frequently after starting; it is contraindicated in POLG mitochondrial disease and children younger than 2 years with clinical suspicion of mitochondrial disease [103]. Lamotrigine (Lamictal XR®) has a warning for serious rashes, including Stevens–Johnson syndrome, with an incidence of approximately 0.8% in children aged 2 to 16 years receiving adjunctive immediate-release formulation and one rash-related death in a prospective cohort of 1983 children; concomitant valproate, exceeding the initial dose or escalation rate and HLA-B*1502 increase risk, almost all cases occur within the first 2 to 8 weeks and the drug should be stopped at the first rash sign [104]. According to GeneGuidance curation of US labels, no valproate label mentions infantile spasms, LGS or lissencephaly, and lamotrigine is approved adjunctively from 2 years for focal, primary generalised tonic–clonic and generalised LGS seizures, so the combination is off-label in this disease and in infants [15]. Sources consulted for this article did not include the valproate and lamotrigine label dosage sections; doses follow the national SmPC and paediatric formularies, with slow lamotrigine titration in the presence of valproate, as required by the boxed warning [104].

7.6. Non-pharmacological epilepsy treatment

Ketogenic diet appears second-line for IESS in both guidelines and among later LGS options in NICE, and was among the best-responding interventions in the Swedish cohort [37, 38, 82]. In a boy with MDS and seizures refractory to valproate, levetiracetam and vigabatrin (up to 94.3 mg/kg/day), an enteral ketogenic diet with commercial formula at a 1.25:1 ratio, delivered through gastrostomy with jejunal extension from 12 months, was associated with seizure freedom until death at 2 years from intestinal volvulus, with weight gain and fewer hospitalisations; this is a single case without a control period [67]. Surgical resection of heterotopic tissue has been attempted in few individuals with SBH, with generally inadequate results, and is not recommended [4, 105]. A systematic review of non-pharmacological SBH treatments in drug-resistant epilepsy found improved seizure severity in 16 of 26 patients, with callosotomy, temporal lobectomy and, less frequently, vagus nerve stimulation or thalamic deep brain stimulation (DBS) [4, 106]. In Lisbon, at Hospital de Santa Maria, the first two SBH patients undergoing bilateral anterior thalamic nucleus DBS had, at 12 and 18 months, more than 50% lower seizure frequency and longer seizure-free periods; both developed postoperative depressive syndrome, responding fully in one and partially in the other to antidepressants [107]. The DCX GeneReview considers, based on these small series, that DBS may improve SBH epilepsy [4]. In MDS with refractory IESS, total callosotomy at 2 years and 4 months eliminated spasm clusters (section 10) [108]. A 2025 review states that callosotomy relieves drug-resistant spasms, and a 2026 report, known only through its title and metadata, describes complete seizure freedom after laser callosotomy in a patient with lissencephaly [64, 109].

7.7. Feeding, breathing and other medical problems

Poor neonatal feeding is managed with a nasogastric tube, usually sufficient in the first weeks; oromotor dysfunction should be assessed at every visit, with clinical swallowing assessment or radiographic investigation for choking, chronic cough, poor weight gain, frequent respiratory infections or unexplained food refusal [3, 4]. When oral intake is safe, feeding therapy (OT or speech and language therapy) includes thickening or cooling foods; severe dysfunction requires a nasogastric tube or gastrostomy, necessary in at least half of children with PAFAH1B1 lissencephaly [3, 4]. Gastrostomy with jejunal extension may be necessary for vomiting and recurrent aspiration, as in the MDS case described in 2026 [67]. Constipation is treated with stool softeners, prokinetics, osmotic agents or laxatives, and the DCX GeneReview considers continuous use safe [3, 4]. Vitamin D supplementation is part of musculoskeletal management when indicated [4]. Respiratory physiotherapy was among the supportive therapies rated most favourably by parents in the German cohort, and the authors attribute improved survival in the current era to aggressive infection management, routine gastrostomy and respiratory physiotherapy [33]. In XLAG, intensive support for neonatal seizures, temperature dysregulation, chronic diarrhoea and nutrition is central [15].

7.8. Motor function, spasticity and rehabilitation

Physiotherapy aims to maximise mobility and reduce late orthopaedic complications (contractures, scoliosis, hip dislocation), with stretching, positioning and adapted equipment (wheelchairs, walking frames, bath chairs, orthoses, adapted pushchairs); OT addresses fine motor skills and oromotor control [3, 4]. Hypertonia and dystonia may be treated by specialists with baclofen, tizanidine, botulinum toxin (Botox®), antiparkinsonian drugs or orthopaedic procedures; the consulted sources provide no doses for this population [3, 4]. For dystonia in Partington syndrome, an ARX allelic disorder, a systematic review found mild improvement only with levodopa and baclofen [15]. Orthopaedic correction of scoliosis and major joint dislocations forms part of musculoskeletal management [4]. In the German cohort, parents rated physiotherapy (23 of 25) and respiratory physiotherapy (10 of 11) good or excellent, and hippotherapy, hydrotherapy, vibration therapy, the Bobath approach and OT approximately 70% effective; early visual stimulation and craniosacral therapy had no or partial effect in approximately half; this is a subjective parental assessment without a control group [33].

7.9. Development, communication and behaviour

In the USA, early intervention from 0 to 3 years, developmental preschool from 3 to 5 years and individualised education plans constitute the GeneReviews model, with variation by country [3, 4]. Augmentative and alternative communication, assessed by a speech and language therapist and ranging from picture exchange to voice-generating devices, supports rather than impairs speech development [4]. In autism spectrum disorder, interventions include applied behaviour analysis; the developmental paediatrician guides strategies and medication, for example for attention deficit hyperactivity disorder, and the child psychiatrist addresses severe aggression; sedating drugs should be weighed against effects on education, therapies, sleep and quality of life [4]. Cerebral visual impairment benefits from psychoeducational activities and early intervention programmes [4].

7.10. Treatment of allelic disorders and specific forms

There is no targeted therapy for XLAG or other ARX phenotypes; in a family with a nuclear localisation signal variant, seizures responded completely but transiently to pyridoxal phosphate and were then controlled with valproate, an isolated report that does not constitute a recommendation [15]. In VLDLR cerebellar hypoplasia, seizures and strabismus receive standard treatment, with early intervention [6]. In tubulinopathies with CFEOM, non-surgical or surgical ophthalmological treatment may be necessary [5].

In LIS9 caused by MACF1 there is no targeted treatment and no registered clinical trial; a search for “MACF1” on ClinicalTrials.gov in November 2026 returned no studies, and the gain of function demonstrated in 2025 has not yet translated into a therapeutic strategy [9, 110]. Epilepsy, present in 89% and starting in the first year in 83%, with infantile spasms in half, is treated according to the general guidelines for infantile epileptic spasms and drug-resistant epilepsy, with the drugs and doses given in sections 7.3 to 7.5; there are no published data on drug-specific response in this entity, apart from resolution of spasms with corticosteroid therapy in one child of the founding series [8, 9, 37, 38]. Dysphagia (four of eight, two with gastrostomy), cortical visual impairment, strabismus, involuntary movements and the neurogenic bladder and bowel dysfunction described in one patient guide swallowing assessment, ophthalmology, motor and communication rehabilitation (including picture exchange communication systems) and urological assessment, according to the general principles for the lissencephalies [3, 8].

In LIS10 caused by CEP85L there is likewise no targeted treatment and no registered trial (a search for “CEP85L” on ClinicalTrials.gov in November 2026 found no studies) [110]. Epilepsy, present in approximately 90% and drug-resistant in 9 of 12 patients with seizures in the 2020 series, is treated with the drugs and doses given in sections 7.3 to 7.5, and the cognitive regression observed after refractory seizures justifies early referral to an epilepsy centre [12, 21]. Patients with normal cognition or mild ID reach adulthood and reproductive age, which requires counselling about teratogenic antiseizure medications, and family V of the 2020 series shows the need to monitor mental health [11, 12].

7.11. Regulatory status

No medicine has a Food and Drug Administration (FDA, USA) or European Medicines Agency (EMA) approved indication for lissencephaly or MDS; all pharmacological treatment is symptomatic and directed at epilepsy [100, 102, 110].

National access data (INFARMED in Portugal and ANVISA in Brazil) were absent from the sources consulted for this article.

7.12. Trials and clinical studies

The October 2026 ClinicalTrials.gov search (lissencephaly, Miller–Dieker, SBH, double cortex) returned seven records, none a pharmacological or gene therapy trial targeting lissencephaly or MDS; other databases (European Union Clinical Trials Register, EU CTR, and World Health Organization International Clinical Trials Registry Platform, ICTRP) were not consulted, and a negative search does not prove absence of studies [110]. Available evidence is observational or derives from trials in broader populations.

These studies do not allow indirect comparisons: they differ in population (PAFAH1B1 alone, several aetiologies or SBH alone), response definition and observation time, and efficacy assessment in retrospective series depends on clinical records or parental perception, with selection and recall bias [33, 39, 81]. A “good or partial response” proportion does not equal seizure freedom, and in the 2024 cohort only 20% were seizure-free in the late phase [81]. Absence of published negative trials reflects absence of trials, not proven efficacy of options in use.

7.13. Developments in the past 12 months and research

The two organoid studies diverge: in MDS and PIDD1 deficiency, the mTOR pathway is hypoactive and phenotype is rescued by an mTORC1 activator, whereas in PAFAH1B1 lissencephaly no primary mTOR dysregulation was observed and an mTORC1 inhibitor partly corrected premature neurogenesis and proteotoxic stress [72, 74]. Neither justifies clinical use, and both contrast with type II focal cortical dysplasia, tuberous sclerosis and hemimegalencephaly, in which mTOR is hyperactivated [72, 74]. A 2025 review proposes speculative MDS therapeutic alternatives (baclofen, ISRIB, NRF2 activators, JAK inhibitors, resveratrol, minocycline, betaine, CRISPR gene editing) without patient data; the same review contains unsupported statements, such as frequent intrauterine death at 10–20 weeks, which should not be reproduced [64]. No application, Committee for Medicinal Products for Human Use (CHMP) opinion or authorisation specific to lissencephaly or MDS was identified in the preceding 12 months, although a negative search does not prove absence [110].

7.14. Agents and situations to avoid

The DCX GeneReview identifies no drugs to avoid specifically, but recommends exploring and avoiding seizure triggers [4]. In LGS, drugs NICE considers potentially aggravating should be used with response monitoring (section 7.5) [38]. Valproate is contraindicated in POLG mitochondrial disease and should be avoided in polytherapy below 2 years whenever possible [103]. Vigabatrin should not be combined with other retinotoxic drugs [100]. Vigabatrin plus hormonal treatment requires caution because of reported movement disorders, encephalopathy and dysautonomia [37]. Therapies unvalidated by a randomised trial should not be used first-line in IESS except within trials [37].

7.15. Anaesthesia and perioperative period

MDS perioperative recommendations are based on three published cases and represent author opinion: continue antiepileptics, including on the morning of surgery, with intravenous alternatives if necessary; prepare for a difficult airway because of micrognathia; obtain pre-anaesthetic echocardiogram and electrocardiogram because of heart disease; do not rely on the bispectral index (BIS), with low baseline values in this syndrome; prefer short-acting agents; avoid suxamethonium in hypotonia because of theoretical rhabdomyolysis and hyperkalaemia risk; monitor train-of-four (TOF) because of possible exaggerated response to non-depolarising blockers; use ultrasound-guided vascular access and prevent hypothermia; and perform sacral dimple ultrasonography before neuraxial anaesthesia [65]. In the described case, a 3-year-old child with MDS undergoing cystoscopy and lithotripsy, induction was inhalational with sevoflurane maintaining spontaneous ventilation until mask ventilation was confirmed, videolaryngoscopy showed a grade I view, and bronchospasm on emergence responded to adrenaline (1 µg/kg); discharge occurred the same day [65].

7.16. Pregnancy, lactation and transition

Women with DCX SBH and seizures should receive close follow-up during pregnancy from a neurologist experienced in epilepsy, preferably at an epilepsy centre; they should be counselled about the teratogenic risk of their current antiepileptic, because some drugs and combinations substantially increase malformation risk, and encouraged to maintain seizure control while understanding withdrawal risks; selection or substitution of more teratogenic drugs should be discussed before planned pregnancy, and the neonate may require measures because of fetal exposure [4]. The US valproate label warns of neural tube defects, reduced IQ and neurodevelopmental disorders after fetal exposure, and NICE links topiramate to a pregnancy prevention programme [38, 103]. The consulted sources do not specifically address lactation in the disorders covered by this article. Transition to adult care should be planned from 10 years, according to the American Epilepsy Society document cited by the DCX GeneReview [4].

7.17. Palliative care and family support

Ongoing assessment of palliative care and home nursing needs, caregiver respite, coordination of appointments, equipment and medication, and social service support form part of management [3, 4]. In the 2026 MDS case, paediatric palliative care was introduced early [67]. Family quality of life measured by PedsQL in 18 families was significantly lower than in families without chronic disease, with the worst domains in worries and daily activities, supporting structured psychosocial support [33].

8. Reproductive Options and Family Counselling

8.1. PAFAH1B1 lissencephaly and SBH

Inheritance is autosomal dominant and the alteration typically de novo, whether an intragenic variant or, rarely, a chromosomal rearrangement disrupting the gene [3]. Vertical maternal transmission has been described in three unrelated families: a boy with posterior ILS and epilepsy inheriting a missense variant from a similarly affected mother; two girls with severe ID, epilepsy, anterior-predominant lissencephaly and basal ganglia calcifications, born to a mother with the same variant and anterior lissencephaly; and a boy with ILS/SBH inheriting a missense variant from a mosaic mother [3]. Parents should be tested for the proband's variant and undergo karyotyping if it is a rearrangement [3]. If the variant is absent from the parents, it may be de novo, confirmed only by parental identity testing, or result from parental germline (or somatic and germline) mosaicism; leucocytes may fail to detect somatic mosaicism and do not detect exclusively germline mosaicism, and a mosaic parent may be mildly affected [3]. Sibling risk is 50% if a parent is a carrier, increased and rearrangement-dependent if a parent has a structural rearrangement, and slightly above general population risk if the variant is absent from parental leucocytes; an affected individual's offspring have a 50% risk [3]. In the series of 811 patients, parental gonadal mosaicism was recognised in several lissencephaly genes, including two siblings with the same TUBG1 variant and blood-negative parents [7].

8.2. Miller–Dieker syndrome

Most deletions arise de novo, during gametogenesis or early embryogenesis, and recurrence risk with normal parental chromosomes is very low, below 1% [16, 59]. However, deletion may result from a familial rearrangement, and karyotype and FISH in both parents are always indicated; de novo abnormalities such as ring chromosome 17 or terminal deletion have negligible recurrence risk, whereas risk may be high in families with balanced rearrangements [34, 59]. Sources diverge on the proportion of MDS cases caused by parental rearrangement: the Oxford Desk Reference cites approximately 12% and the Brazilian book approximately 20%, the latter adding a 20% pregnancy-loss probability in these cases [16, 59]. The figures appear to originate from Pollin and colleagues' 1999 study, which measures something else: in 14 families with a reciprocal translocation between 17p13 and the distal region of another autosome, ascertained through an index case, 38 balanced carriers had 127 pregnancies, with abnormal outcomes in 33 (26%), including 15 (12%) with MDS due to del(17p), 9 (7%) with a milder dup(17p) phenotype and 9 unstudied; excluding unexplained losses, abnormal outcomes reached 33 of 99 pregnancies (33%) [20]. There were also 26 miscarriages (20%) and 2 stillbirths (2%), frequencies similar to the general population (10–20% and 1%) [20]. Thus, 12% corresponds to empirical MDS risk per pregnancy of a balanced carrier, and 20% to the miscarriage rate in those pregnancies; the online sources read do not provide the proportion of MDS cases due to parental rearrangement, and the 12% and 20% presented as such by the two books should be read with this qualification [16, 20, 59]. In an Italian case, an unbalanced t(1;17)(q44;p13.2) translocation arose de novo, showing that not every unbalanced translocation is inherited [114]. When a parent is a carrier, prenatal diagnosis and PGT are possible [34].

8.3. DCX disorders

The father of an affected boy is not a carrier and need not be tested [4]. Mothers of more than half of males with DCX variants are clinically asymptomatic, and asymptomatic heterozygotes without evident structural abnormalities have been described [4]. A mother with more than one affected child, no other affected relatives and no variant detected in leucocytes probably has gonadal mosaicism; preliminary data indicate that up to 10% of unaffected mothers of probands with a DCX variant may have gonadal mosaicism, with or without somatic mosaicism [4, 115]. The mother should be tested in leucocytes and, if negative, other tissues (hair roots, buccal mucosa), and undergo neurological examination and MRI to investigate SBH [4]. Parents of approximately 90% of female probands are clinically unaffected [4]. Paternal transmission of SBH through somatic and gonadal DCX mosaicism in a male was described in work from the Centro de Genética Preditiva e Preventiva in Porto and Hospital de Santo António, requiring consideration of paternal testing with sufficient mosaicism sensitivity when a female proband's variant is absent from her mother [4, 116].

When the mother is heterozygous, the maternal grandmother and aunts are at risk of being carriers, and testing at-risk women is most informative when the proband's variant is known [4]. X-inactivation pattern analysis in any prenatal or postnatal sample has no value for predicting manifestations [4]. Carrier detection uses DCX testing in mothers of patients with an identified variant; without an identified variant, maternal MRI may be used [16].

8.4. Tubulinopathies, ARX and recessive genes

In tubulinopathies, inheritance is autosomal dominant, more than 95% of cases are de novo and a parent is rarely affected, usually with TUBB3 or, less frequently, TUBB2B [5]. Maternal germline mosaicism has been described in two families with several affected children, and somatic mosaics may be mildly affected; sibling risk is slightly above population risk if the variant is absent from the parents, and offspring risk is 50% [5]. In XLAG, females heterozygous for ARX loss of function may be asymptomatic or have ID and epilepsy, and counselling follows X-linked inheritance [3, 80]. In VLDLR cerebellar hypoplasia and RELN LCH, autosomal recessive inheritance gives siblings a 25% risk of being affected, 50% of being asymptomatic carriers and 25% of being neither affected nor carriers; carrier testing, prenatal diagnosis and PGT are possible when variants are known [6, 16]. In dystroglycanopathies with cobblestone malformation, recurrence risk is 25% [2, 16].

In LIS9 caused by MACF1, all alterations in the founding series were described as de novo, and in the 2025 cohort the de novo status of the new zinc residue substitutions was confirmed in three individuals [8, 9]. For siblings, the risk is low but not zero, because of the possibility of germline mosaicism, a principle applied to de novo dominant disorders such as the tubulinopathies, without empirical data specific to LIS9; for the offspring of an affected individual, the theoretical risk is 50%, with no reproduction described [5, 9]. The biallelic form described, in a child of consanguineous heterozygous parents, implies a 25% risk in each pregnancy [9]. Prenatal or preimplantation diagnosis is possible when the familial variant is known; fetal ultrasound and MRI have no validated criteria for LIS9 in the sources consulted [2].

In LIS10 caused by CEP85L, inheritance is autosomal dominant: in the 2020 series, the variants were de novo in four patients, inherited in two families with a mild phenotype (one with at least five affected across three generations) and present in two sisters without the variant in the parents, through probable gonadal mosaicism, with unknown origin in two [11, 12]. In the 2025 review, 9 variants were de novo, 10 inherited from a parent and 13 of unknown origin [21]. The risk for the offspring of an affected individual is 50%, with variable expressivity within the same family (pachygyria in one member and SBH in others), and the parents of an apparently de novo case should be examined by MRI and tested, because a parent with mild SBH may go unnoticed and gonadal mosaicism has documented recurrence [12, 79]. Prenatal diagnosis has already been described [21].

8.5. Counselling without a molecular diagnosis

The table follows the Neuro-MIG consensus, which warns that empirical counselling requires very high confidence in MRI interpretation [2]. Even with negative testing, parents of children with thin (undulating) lissencephaly should be counselled about possible autosomal recessive inheritance; SBH (particularly in females), thick frontal lissencephaly and the temporal XLAG pattern predict X-linked inheritance [7]. For ILS without a genetic diagnosis, the Oxford Desk Reference estimates a pregenomic recurrence risk of approximately 10%, potentially reaching 25%, particularly with microcephaly, and recommends cautious use of these figures; in undiagnosed neonatal encephalopathy, it considers recurrence risk up to 25% [16, 84]. DNA storage is recommended in probands without a molecular diagnosis [3].

8.6. Prenatal diagnosis, preimplantation testing and other issues

The ideal time to discuss genetic risk and prenatal and PGT options is before pregnancy, and counselling should be offered to affected or at-risk young adults [3, 4]. With a known familial variant, prenatal diagnosis can be performed by chorionic villus sampling at 11 weeks [16]. Fetal imaging does not exclude recurrence because of limitations described in section 4 [3, 16]. In a female fetus with a DCX variant, prenatal testing should be offered only after discussion of reduced penetrance and variability [4]. Perspectives on prenatal testing differ between families and professionals, and the decision is personal [3]. Testing asymptomatic minors is not specifically addressed in the consulted sources for the disorders covered by this article; in DCX families, carrier variability and lack of predictive value of X-inactivation studies are considerations in that decision [4]. In Hutterite communities, the VLDLR deletion carrier frequency (1 in 15) may justify population reproductive screening [6].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and what each figure measures

Prevalence figures are mutually consistent but their provenance is weak, and no PAFAH1B1-specific rate exists [15]. Relative gene frequency, conversely, rests on large series: in lissencephaly generally, PAFAH1B1 accounts for approximately 40% and DCX 23% of cases, and MDS 9% (series of 811 patients) [2, 7]. Approximately 80% of ILS patients have a PAFAH1B1 deletion or variant, and DCX explains approximately 12% (boys) according to the Oxford Desk Reference [16]. DCX explains virtually all families with X-linked lissencephaly or SBH, approximately 10% of all thick classical lissencephaly (38% in males), 53–85% of all SBH, up to 88% of sporadic SBH and approximately 29% of male SBH [4]. Female SBH is probably under-recognised because mildly affected individuals may never undergo MRI [15]. Nearly 200 individuals with an intragenic PAFAH1B1 variant and more than 300 with a DCX variant had been reported to date [3, 4].

9.2. United States and Europe

North American data derive mainly from the Seattle cohort, which assembled approximately 1400 individuals with lissencephaly over 30 years, and the Atlanta programme [3, 7]. Comparison of the Seattle cohort (129 patients, approximately 80% referred for imaging review, biased towards severe forms) and Florence cohort (59 patients, a European epilepsy centre) showed more partial and thin SBH in Florence and no clinically severe lissencephaly, probably because of early lethality, illustrating how recruitment location alters observed distribution; molecular resolution rates nevertheless overlapped (84% and 76%) [7, 58]. In Europe, besides Florence, there are German, Austrian and Swiss series (47 patients, recruited through a patient association), a Swedish series (20 epilepsy patients) and the Leger and Bahi-Buisson DCX cohorts [4, 33, 82]. None measures prevalence.

9.3. Founder populations and other regions

The clearest example of a founder effect is the homozygous 199 kb VLDLR deletion in the Hutterites of Canada and the USA, present in all three branches of the community (Schmiedeleut, Lehrerleut and Dariusleut), with more than 25 individuals followed for years and a carrier frequency estimated at 1 in 15 [6, 55]. Outside the Hutterites, VLDLR cerebellar hypoplasia has been described in consanguineous families, notably from Turkey, where some patients developed quadrupedal locomotion [6]. In the differential, Fukuyama congenital muscular dystrophy is relatively common in Japan, owing to a 3 kb founder insertion in FKTN, and muscle-eye-brain disease in Finland [16]. PIDD1 deficiency was identified in consanguineous families from the Yale neurogenetics cohort [72]. In China, a 2026 cohort of 40 patients identified PAFAH1B1 point variants and 17p13.3 microdeletions as the predominant causes (31.3% in the prenatal group and 25.5% in the postnatal group; denominators per group not given), and a synthesis of 210 literature cases attributed 26% to PAFAH1B1, 12% to DYNC1H1 and 10% to DCX [73]. In Latin America, the 2020 consensus cites an Argentine cohort of 38 patients with lissencephaly, SBH or periventricular nodular heterotopia with a 36% yield, and a Uruguayan case of PAFAH1B1 lissencephaly was diagnosed within the Urugenomes genome project [2, 90].

LIS9 has no known founder effect: the eight patients of the founding series came from Mumbai, Rotterdam, California (including twin girls of Hispanic origin), Warsaw, London (a Syrian family) and Saitama, with no geographical clustering of variants [8]. The only biallelic form described occurred in a consanguineous family [9]. LIS10 likewise has no known founder effect: it has been described in Taiwan, Belgium, the USA, Ireland, Canada, Italy and Germany [11, 12, 21, 61, 78, 79, 117].

9.4. Portugal and Brazil

In the consulted sources, Portuguese data are clinical rather than epidemiological: paternal SBH transmission through somatic DCX mosaicism (Porto) and the first anterior thalamic nucleus DBS series in SBH (Lisbon), both cited by the DCX GeneReview; Lisbon also hosted a Neuro-MIG consensus meeting in 2018 [1, 107, 116]. In Brazil, a Universidade Estadual de Campinas (UNICAMP) series studied 15 patients with classical lissencephaly or SBH and described a LIS1 variant with a more severe phenotype than expected for a missense variant [118]. The Brazilian book Síndromes Genéticas classifies MDS as rare with unknown prevalence and identifies the Centro de Pesquisa sobre o Genoma Humano e Células-Tronco at the Universidade de São Paulo as a resource [59]. The consulted sources contain no prevalence estimates for Portugal or Brazil, which does not mean no published data exist.

9.5. Special populations

Women with DCX SBH reach adulthood and reproductive age, sometimes undiagnosed until an affected child is born, and require counselling about teratogenic antiepileptics and the 50% offspring risk [4]. Adults with lissencephaly are increasingly numerous: in the German cohort, 18% of PAFAH1B1 patients were older than 15 years and the oldest was 25, while in Leger's series ten DCX males were adults, making adult care transition a practical necessity [4, 33]. 46,XY neonates with XLAG need a sex development team [15]. Prenatal exposures such as drug use may obscure a genetic cause, as in the Uruguayan case [90].

9.6. European Reference Networks and patient associations

ERN EpiCARE, the European Reference Network for rare and complex epilepsies, offers a patient registry, working groups and a case discussion platform (CPMS) relevant to drug-resistant epilepsy in lissencephaly and MDS [40]. ERN ITHACA covers rare congenital malformations and intellectual disability, including MDS and malformation syndromes with lissencephaly [41]. Neuro-MIG (COST action CA16118) brings together European researchers in malformations of cortical development and authored the 2020 consensus statements [1, 42]. In Germany, the LISS e.V. parent association (Hilfeforum Lissenzephalie) provides information, regional contacts, bereavement support and annual meetings, and helped recruit the 2024–2026 cohort [33, 119]. In the USA, GeneReviews lists the National Institute of Neurological Disorders and Stroke (NINDS) lissencephaly information page, American Epilepsy Society, Epilepsy Foundation and American Association on Intellectual and Developmental Disabilities (AAIDD) [3–5]. MedlinePlus Genetics maintains an MDS page linking to the National Organization for Rare Disorders (NORD) [120]. The consulted sources identified no specific lissencephaly associations in Portugal or Brazil.

10. Illustrative Clinical Case

10.1. Miller–Dieker syndrome with refractory spasms treated by total callosotomy

Fukuoka and colleagues published in 2024 the case of a boy born at 30 weeks by emergency caesarean section for non-reassuring fetal status, weighing 975 g and without relevant family history [108]. He had unusual facies, with a small head, jaw and nose and low-set ears. MRI at 76 days showed severe lissencephaly, and comparative genomic hybridisation identified a 17p13.3-p13.2 deletion including LIS1 (PAFAH1B1) and YWHAE, confirming MDS [108]. At 6 months, focal seizures with impaired awareness, apnoea and behavioural arrest developed and were controlled with phenobarbital. At 9 months, spasm clusters appeared, with interictal EEG hypsarrhythmia and high-amplitude generalised irregular slow waves ictally, constituting an infantile spasm syndrome [108]. Valproate, topiramate and levetiracetam were ineffective, and ACTH and ketogenic diet failed to stop spasms or hypsarrhythmia; vigabatrin was not approved in Japan at the time and was not used [108]. At 10 months, tube feeding began and repeated aspiration pneumonias led to tracheostomy; at one year he had neither head control nor verbal comprehension [108].

At 2 years, he still had 4 to 5 spasm clusters daily despite clobazam, zonisamide and rufinamide. MRI showed a four-layer cortex with a cell-sparse layer and absent gyri, and a short thick corpus callosum; iomazenil single-photon emission computed tomography showed no focal deficit, and EEG showed bilateral diffuse high-voltage spike–wave discharges [108]. Because resection was considered unhelpful, the team performed palliative total callosotomy at 2 years and 4 months. Intraoperative EEG showed near-continuous bilaterally synchronous discharges disappearing immediately after disconnection, with persistent asynchronous multifocal spikes [108]. Postoperatively, spasm clusters disappeared, replaced by isolated spasms followed by focal tonic seizures, without new functional deficit; the child smiled more and responded better to stimuli. Three years later, seizures occurred a few times daily and clusters were rare; developmental indicators showed no measurable gains, but positive effects on smiling and responsiveness persisted and parental satisfaction was good [108].

Article Figure 1, whose legend was consulted, shows T2-weighted MRI before surgery (axial sections with a four-layer cortex, cell-sparse layer and absent gyri; sagittal section with a short thick corpus callosum) and after surgery (completely sectioned corpus callosum): https://pmc.ncbi.nlm.nih.gov/articles/PMC11079450/figure/f0005/ [108]. The authors present this as the first report of total callosotomy in MDS-associated IESS. It is a single case and callosotomy is not routine management: IESS guidelines place hormonal treatment and vigabatrin first-line and reserve surgery for non-responders, particularly with a focal lesion, a condition this patient did not meet [37, 38]. The case illustrates the typical MDS sequence (severe lissencephaly with dysmorphism, focal seizures followed by refractory spasms, dysphagia with aspiration) and shows that a realistic goal in diffuse lissencephaly may be reducing seizure burden and improving interaction without developmental gains.

10.2. Additional cases

A 2018 Italian case describes a boy with ventriculomegaly, corpus callosum agenesis and growth restriction on second-trimester ultrasonography, born at 37 weeks weighing 1800 g, with generalised agyria on MRI at 5 days and first-year electroclinical evolution from tonic movements at 4 months to “subtle spasms” with modified hypsarrhythmia at 6 months, typical clustered spasms at 7 months and classical hypsarrhythmia at 8 months; CMA showed an approximately 3.8 Mb 17p13.3p13.2 deletion with an approximately 4.2 Mb 1q44 duplication, and FISH a de novo unbalanced t(1;17)(q44;p13.2) translocation [114]. Figure 1 of that article, with MRI and EEG at 5 days, is at https://pmc.ncbi.nlm.nih.gov/articles/PMC6144613/figure/F1/ (legend consulted) [114]. A 2022 Uruguayan case in a 7-year-old boy with microcephaly from birth, bilateral spastic cerebral palsy, profound ID and drug-resistant epilepsy, whose pregnancy involved multiple drug use, was resolved only by genome sequencing, identifying PAFAH1B1 NM_000430.4:c.681dup p.(Lys228GlufsTer28), absent from his mother and classified pathogenic [13, 90].

A Dutch boy from the 2018 series with the de novo variant p.Cys7135Phe (NM_001394062.1:c.21707G>T, p.(Cys7236Phe)) illustrates LIS9 [8]. He had hypotonia without spasticity and global delay; at 7.5 years he weighed 28 kg (+1 SD), measured 128 cm (0 SD) and had an HC of 53 cm (+0.7 SD), did not sit or walk, said 10 syllables and had severe ID, mixed involuntary movements, eye movement abnormalities and exotropia, and was fed by gastrostomy [8]. Seizures began at 6 months with infantile spasms and evolved to Lennox–Gastaut syndrome [8]. MRI showed posterior-predominant diffuse pachygyria with mildly thickened cortex, dysplastic hippocampi, a thin anterior commissure and corpus callosum, a very narrow and broad pons with a diminutive basis and mild hypoplasia of the vermis and cerebellar hemispheres [8]. Figure 3 of the article (https://pmc.ncbi.nlm.nih.gov/articles/PMC6288423/), whose legend was read, compares this patient's DTI with that of a control: small midbrain with reduced decussation of the superior cerebellar peduncles, median ventral cleft in the upper pons, absence of the anterior and posterior transverse pontine fibres and emergence of the trigeminal nerves at the extreme lateral surface of the pons [8]. The Japanese patient of the same series was published in 2014 as Irahara syndrome [60].

11. Prognosis and Long-Term Follow-up

11.1. Survival and course by entity

Life expectancy in lissencephaly of any cause relates to imaging malformation severity [3, 121]. On Di Donato's scale, severe lissencephaly (grade 3) has approximately 50% mortality by 10 years with a normal cerebellum and much higher mortality with cerebellar hypoplasia, while survival into adulthood is expected or frequent in grades 1 and 2 [58]. In PAFAH1B1 ILS, approximately 50% of patients live to 10 years and very few reach 20; death is usually due to pneumonia or status epilepticus, and the oldest known individual with complete lissencephaly lived to 30 years, although a mildly affected patient with grade 4b lissencephaly was alive at 49 years [3]. In the German–Austrian–Swiss cohort, 5 of 38 PAFAH1B1 patients died (median age at death 9.1 years, cause unknown), the oldest living patient was 25 and 7 (18%) were older than 15 years; one of the 5 DCX males died at 6.8 years [33]. The authors consider that, compared with 1991–2010 series, survival has increased in the current era through aggressive infection management, routine gastrostomy and respiratory physiotherapy, with a significant proportion reaching 20 years or older [33]. In de Wit's series cited in a 2018 report, 11 of 24 patients with type 1 lissencephaly were alive after mean follow-up of 14 years [114, 121].

In MDS, developmental prognosis is poor in all patients, many die within the first two years and few reach 10 years; the oldest known individual died at 17 years [3]. Aspiration pneumonia due to poor airway control is the most common cause of death [3, 59]. Males with thick DCX lissencephaly may survive into adulthood (ten adults up to 37 years in Leger's series of 33), but life expectancy is generally shortened by seizure complications, deficient central regulation of vital functions such as breathing, and aspiration [4]. Women with SBH generally live into adulthood, without reliable life expectancy data, probably shortened with severe ID or intractable epilepsy [3]. In tubulinopathies, mild forms survive into adulthood and severe forms may die early from seizures or pneumonia [5]. In VLDLR cerebellar hypoplasia, Hutterite experience suggests no significant reduction in life expectancy [6]. In XLAG, lethality within the first three months is high [4].

Development depends mainly on malformation severity and epilepsy control. In complete lissencephaly, the maximum level is approximately 3 to 5 months, and poor seizure control may cause regression to neonatal levels [3]. Regression occurred in six of nine reassessed PAFAH1B1 patients in the German cohort, associated with worsening epilepsy and nutritional status, and patients without spasms attained more milestones [33, 81]. Males with DCX attain more milestones than PAFAH1B1 patients, and almost half walk independently [4, 33].

In LIS9 caused by MACF1 there are no survival studies; the patients of the founding series were aged between 2.5 and 16 years at last assessment, and in the 2025 cohort one patient with an EF-hand variant underwent MRI at 52 years, which shows that survival into adulthood is possible [8, 9]. The functional prognosis is guarded: severe ID in all those assessed, inability to sit in 38% and to walk in 47%, absence of language in 56% and epilepsy in 89% [9]. Limited ocular abduction may improve, as in the Irahara patient, whose abduction normalised between 2.5 and 5 years [8]. Absence of data on life expectancy does not equate to normal life expectancy.

In LIS10 caused by CEP85L, most patients of the 2020 series were adults (up to 52 years), 12 of 13 walked and all spoke, and the prognosis depends mainly on control of epilepsy, drug-resistant in most, and on the severity of the malformation, which ranges from mild SBH to posterior agyria [12]. There are no survival studies, and the advanced age of several patients does not demonstrate normal life expectancy [12, 21].

11.2. Surveillance

12. Summary and Final Considerations

Lissencephalies form a continuous spectrum from diffuse agyria to subcortical band heterotopia, unified by failed neuronal migration and now classified by degree, gradient, cortical thickness and associated malformations rather than former “type I” and “type II” designations [1, 58]. The malformation is considered exclusively genetic, and combined CMA and NGS investigation solves more than 80% of cases; four genes, PAFAH1B1, DCX, TUBA1A and DYNC1H1, explain approximately 70% [2, 7]. Specialist MRI interpretation guides testing: posterior agyria points to PAFAH1B1, anterior agyria and SBH to DCX, basal ganglia and brainstem malformations to tubulins, thin lissencephaly with cerebellar hypoplasia to the reelin pathway, and temporal lissencephaly with corpus callosum agenesis in a boy with abnormal genitalia to ARX [2, 58]. Miller–Dieker syndrome is the most severe PAFAH1B1 spectrum form, because the 17p13.3 deletion extends to YWHAE and frequently CRK, adding characteristic facies and other organ malformations to agyria; CMA distinguishes it from isolated lissencephaly and distal deletions with YWHAE without PAFAH1B1, a different entity without lissencephaly [53, 54, 66].

The clinical course is dominated by epilepsy, occurring in more than 90% of PAFAH1B1 lissencephaly, usually beginning with spasms at approximately 5 to 6 months and evolving to drug-resistant epilepsy with a Lennox–Gastaut pattern, and by feeding difficulties and aspiration, accounting for most mortality [3, 81]. There is no disease-modifying treatment or registered therapeutic trial; spasms are treated according to general IESS guidelines, which diverge between prednisolone plus vigabatrin (NICE) and hormonal treatment alone or combined (GNP), while preference for valproate and lamotrigine in chronic epilepsy rests on retrospective series and must be weighed against hepatotoxicity and serious rash warnings and the list of potentially aggravating drugs in LGS [37–39, 103, 104]. Early gastrostomy, respiratory physiotherapy and aggressive infection management are associated with longer survival than in older series, and Di Donato's clinical scale gives families a realistic prognostic reference, with approximately 50% mortality by 10 years in severe lissencephaly and survival into adulthood in mild forms [33, 58].

Genetic counselling depends on cause: most PAFAH1B1 variants and MDS deletions arise de novo, but parental karyotype and FISH are mandatory in MDS because a balanced translocation carrier has an approximately 12% MDS risk per pregnancy; in DCX disorders, X-linked inheritance, reduced female penetrance and gonadal mosaicism up to 10% require maternal testing in several tissues and MRI [4, 20, 115]. Preclinical organoid research has identified potentially treatable pathways, such as mTOR signalling and proteostasis, but with divergent results and no clinical application [72, 74]. European EpiCARE and ITHACA networks and patient associations offer specialist support and access to registries and research [40, 41, 119].

LIS9 caused by MACF1 adds to the spectrum a posterior-predominant lissencephaly with brainstem malformation due to a midline crossing defect, recognisable by the broad and flat “W-shaped” pons and medulla, the absence of transverse pontine fibres and the thin anterior commissure; it results from de novo gain-of-function missense variants in the EF1-EF2-GAR module, the intragenic deletion may escape sequence analysis, and treatment is that of epilepsy and disability, without targeted therapy [8, 9]. LIS10 caused by CEP85L is a posterior lissencephaly or SBH with an abrupt transition, with frequent epilepsy and highly variable cognition, dominantly transmitted in families with a mild phenotype, and should be part of lissencephaly panels [12, 21].

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. PMID: 32779696. PMC7586092.
2. 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. PMID: 32895508. PMC7790753.
3. 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–. PMID: 20301752. Available from: https://www.ncbi.nlm.nih.gov/books/NBK5189/
4. Geis T, Uyanik G, Aigner L, Couillard-Despres S, Winkler J, Hehr U. DCX-Related Disorders. 2007 Oct 19 [updated 2025 Jun 5]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–. PMID: 20301364. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1185/
5. 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–. PMID: 27010057. Available from: https://www.ncbi.nlm.nih.gov/books/NBK350554/
6. Boycott KM, MacDonald SK, Parboosingh JS. VLDLR Cerebellar Hypoplasia. 2008 Aug 26 [updated 2020 Feb 27]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–. PMID: 20301729. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1874/
7. Di Donato N, Timms AE, Aldinger KA, Mirzaa GM, Bennett JT, Collins S, et al. Analysis of 17 genes detects mutations in 81% of 811 patients with lissencephaly. Genet Med. 2018;20(11):1354-1364. doi:10.1038/gim.2018.8. PMID: 29671837. PMC6195491.
8. Dobyns WB, Aldinger KA, Ishak GE, Mirzaa GM, Timms AE, Grout ME, et al. MACF1 Mutations Encoding Highly Conserved Zinc-Binding Residues of the GAR Domain Cause Defects in Neuronal Migration and Axon Guidance. Am J Hum Genet. 2018;103(6):1009-1021. doi:10.1016/j.ajhg.2018.10.019. PMID: 30471716. PMC6288423.
9. Dekker J, Schot R, Aldinger KA, Everman DB, Washington C, Jones JR, et al. A clinical and genotype-phenotype analysis of MACF1 variants. Am J Hum Genet. 2025;112(10):2363-2380. doi:10.1016/j.ajhg.2025.08.010. PMID: 40925378. PMC12431674.
10. Online Mendelian Inheritance in Man, OMIM. Lissencephaly 9 with complex brainstem malformation; LIS9. MIM 618325. Baltimore (MD): Johns Hopkins University; [accessed 2026-11-07]. Available at: https://www.omim.org/entry/618325
11. Online Mendelian Inheritance in Man, OMIM. Lissencephaly 10; LIS10. MIM 618873. Baltimore (MD): Johns Hopkins University; [accessed 2026-11-07]. Available at: https://www.omim.org/entry/618873
12. Tsai MH, Muir AM, Wang WJ, Kang YN, Yang KC, Chao NH, et al. Pathogenic Variants in CEP85L Cause Sporadic and Familial Posterior Predominant Lissencephaly. Neuron. 2020;106(2):237-245.e8. doi:10.1016/j.neuron.2020.01.027. PMID: 32097630. PMC7357395.
13. 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. PMID: 28967166. PMC5765404.
14. National Center for Biotechnology Information. Gene [Internet]. Bethesda (MD): NCBI; Gene ID 1641 (DCX), 5048 (PAFAH1B1), 7531 (YWHAE) [cited 2026 Oct 9]. Available from: https://www.ncbi.nlm.nih.gov/gene/
15. GeneGuidance. Clinical curation of genes PAFAH1B1, DCX, YWHAE, TUBA1A, ARX, RELN and VLDLR [local database]. Version dated 26 September 2026 [accessed 9 October 2026].
16. Firth HV, Hurst JA. Lissencephaly, polymicrogyria, and neuronal migration disorders. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 216–21.
17. Online Mendelian Inheritance in Man, OMIM. Microtubule-actin cross-linking factor 1; MACF1. MIM 608271. Baltimore (MD): Johns Hopkins University; [accessed 2026-11-07]. Available at: https://www.omim.org/entry/608271
18. Online Mendelian Inheritance in Man, OMIM. CEP85-like protein; CEP85L. MIM 618865. Baltimore (MD): Johns Hopkins University; [accessed 2026-11-07]. Available at: https://www.omim.org/entry/618865
19. Orphanet. Lissencephaly type 1 due to doublecortin gene mutation (ORPHA:2148) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/2148
20. Pollin TI, Dobyns WB, Crowe CA, Ledbetter DH, Bailey-Wilson JE, Smith AC. Risk of abnormal pregnancy outcome in carriers of balanced reciprocal translocations involving the Miller-Dieker syndrome (MDS) critical region in chromosome 17p13.3. Am J Med Genet. 1999;85(4):369-375. doi:10.1002/(sici)1096-8628(19990806)85:4<369::aid-ajmg13>3.0.co;2-l. PMID: 10398263.
21. Schumann I, Abou Jamra R, Jauss RT, Karnstedt M, Specht U, Zacher P, et al. Overview and expansion of CEP85L-associated lissencephaly. Eur J Med Genet. 2025;77:105042. doi:10.1016/j.ejmg.2025.105042. PMID: 40850669.
22. Orphanet. Lissencephaly due to TUBA1A mutation (ORPHA:171680) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/171680
23. Gogate N, Jolly A, Rosenfeld JA, Bahena-Carbajal P, Carrasco D, Bernstein JA, et al. Domain-associated phenotypic patterns with MACF1 variants. Genet Med. 2026:102776. doi:10.1016/j.gim.2026.102776. PMID: 42937053.
24. Orphanet. Posterior-predominant lissencephaly-broad flat pons and medulla-midline crossing defects syndrome (ORPHA:572013) [Internet]. Paris: INSERM; [accessed 2026 Nov 7]. Available at: https://www.orpha.net/en/disease/detail/572013
25. Orphanet. Lissencephaly (ORPHA:48471) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/48471
26. Orphanet. Lissencephaly due to LIS1 mutation (ORPHA:95232) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/95232
27. Orphanet. Miller-Dieker syndrome (ORPHA:531) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/531
28. Orphanet. Subcortical band heterotopia (ORPHA:99796) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/99796
29. Orphanet. X-linked lissencephaly with abnormal genitalia (ORPHA:452) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/452
30. Orphanet. Isolated lissencephaly type 1 without known genetic defects (ORPHA:1084) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/1084
31. Orphanet. Distal 17p13.3 microdeletion syndrome (ORPHA:261257) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/261257
32. Orphanet. 17p13.3 microduplication syndrome (ORPHA:217385) [Internet]. Paris: INSERM; [accessed 2026 Oct 9]. Available at: https://www.orpha.net/en/disease/detail/217385
33. Proepper CR, Schwarz LM, Schuetz SM, von Au K, Bast T, Beaud N, et al. Gene-specific long-term course, neurodevelopmental outcome and quality of life in patients with LIS1/PAFAH1B1-, DCX-, DYNC1H1-, TUBA1A- and TUBG1-related lissencephaly. Orphanet J Rare Dis. 2026;21(1):206. doi:10.1186/s13023-026-04398-z. PMID: 42177523. PMC13202861.
34. 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. 258–61.
35. Dobyns WB, Curry CJ, Hoyme HE, Turlington L, Ledbetter DH. Clinical and molecular diagnosis of Miller-Dieker syndrome. Am J Hum Genet. 1991;48(3):584-594. PMID: 1671808. PMC1682996.
36. Koenig M, Dobyns WB, Di Donato N. Lissencephaly: Update on diagnostics and clinical management. Eur J Paediatr Neurol. 2021;35:147-152. doi:10.1016/j.ejpn.2021.09.013. PMID: 34731701.
37. Ramantani G, Bölsterli BK, Alber M, Klepper J, Korinthenberg R, Kurlemann G, et al. Treatment of Infantile Spasm Syndrome: Update from the Interdisciplinary Guideline Committee Coordinated by the German-Speaking Society of Neuropediatrics. Neuropediatrics. 2022;53(6):389-401. doi:10.1055/a-1909-2977. PMID: 35882373. PMC9643068.
38. National Institute for Health and Care Excellence. Epilepsies in children, young people and adults. NICE guideline NG217 [Internet]. London: NICE; 2022 Apr 27 [updated 2026 Aug 5; cited 2026 Oct 9]. Section 6.2 Lennox–Gastaut syndrome; 6.3 Infantile spasms syndrome. Available from: https://www.nice.org.uk/guidance/ng217/chapter/6-Treating-childhood-onset-epilepsies
39. Herbst SM, Proepper CR, Geis T, Borggraefe I, Hahn A, Debus O, et al. LIS1-associated classic lissencephaly: A retrospective, multicenter survey of the epileptogenic phenotype and response to antiepileptic drugs. Brain Dev. 2016;38(4):399-406. doi:10.1016/j.braindev.2015.10.001. PMID: 26494205.
40. European Reference Network for rare and complex epilepsies (EpiCARE) [Internet]. [cited 2026 Oct 9]. Available from: https://epi-care.eu/
41. European Reference Network on Rare Congenital Malformations and Rare Intellectual Disability (ERN ITHACA) [Internet]. Available from: https://ern-ithaca.eu/
42. European Network on Brain Malformations (Neuro-MIG), COST Action CA16118 [Internet]. Available from: https://www.neuro-mig.org/
43. MILLER JQ. LISSENCEPHALY IN 2 SIBLINGS. Neurology. 1963;13:841-850. doi:10.1212/wnl.13.10.841. PMID: 14066999.
44. Dobyns WB, Stratton RF, Parke JT, Greenberg F, Nussbaum RL, Ledbetter DH. Miller-Dieker syndrome: lissencephaly and monosomy 17p. J Pediatr. 1983;102(4):552-558. doi:10.1016/s0022-3476(83)80183-8. PMID: 6834189.
45. Reiner O, Carrozzo R, Shen Y, Wehnert M, Faustinella F, Dobyns WB, et al. Isolation of a Miller-Dieker lissencephaly gene containing G protein beta-subunit-like repeats. Nature. 1993;364(6439):717-721. doi:10.1038/364717a0. PMID: 8355785.
46. Dobyns WB, Reiner O, Carrozzo R, Ledbetter DH. Lissencephaly. A human brain malformation associated with deletion of the LIS1 gene located at chromosome 17p13. JAMA. 1993;270(23):2838-2842. doi:10.1001/jama.270.23.2838. PMID: 7907669.
47. Hattori M, Adachi H, Tsujimoto M, Arai H, Inoue K. Miller-Dieker lissencephaly gene encodes a subunit of brain platelet-activating factor acetylhydrolase. Nature. 1994;370(6486):216-8. doi:10.1038/370216a0. PMID: 8028668.
48. des Portes V, Pinard JM, Billuart P, Vinet MC, Koulakoff A, Carrié A, et al. A novel CNS gene required for neuronal migration and involved in X-linked subcortical laminar heterotopia and lissencephaly syndrome. Cell. 1998;92(1):51-61. doi:10.1016/s0092-8674(00)80898-3. PMID: 9489699.
49. Gleeson JG, Allen KM, Fox JW, Lamperti ED, Berkovic S, Scheffer I, et al. Doublecortin, a brain-specific gene mutated in human X-linked lissencephaly and double cortex syndrome, encodes a putative signaling protein. Cell. 1998;92(1):63-72. doi:10.1016/s0092-8674(00)80899-5. PMID: 9489700.
50. Hirotsune S, Fleck MW, Gambello MJ, Bix GJ, Chen A, Clark GD, et al. Graded reduction of Pafah1b1 (Lis1) activity results in neuronal migration defects and early embryonic lethality. Nat Genet. 1998;19(4):333-339. doi:10.1038/1221. PMID: 9697693.
51. Hong SE, Shugart YY, Huang DT, Shahwan SA, Grant PE, Hourihane JO, et al. Autosomal recessive lissencephaly with cerebellar hypoplasia is associated with human RELN mutations. Nat Genet. 2000;26(1):93-96. doi:10.1038/79246. PMID: 10973257.
52. Kitamura K, Yanazawa M, Sugiyama N, Miura H, Iizuka-Kogo A, Kusaka M, et al. Mutation of ARX causes abnormal development of forebrain and testes in mice and X-linked lissencephaly with abnormal genitalia in humans. Nat Genet. 2002;32(3):359-369. doi:10.1038/ng1009. PMID: 12379852.
53. Toyo-oka K, Shionoya A, Gambello MJ, Cardoso C, Leventer R, Ward HL, et al. 14-3-3epsilon is important for neuronal migration by binding to NUDEL: a molecular explanation for Miller-Dieker syndrome. Nat Genet. 2003;34(3):274-285. doi:10.1038/ng1169. PMID: 12796778.
54. Cardoso C, Leventer RJ, Ward HL, Toyo-Oka K, Chung J, Gross A, et al. Refinement of a 400-kb critical region allows genotypic differentiation between isolated lissencephaly, Miller-Dieker syndrome, and other phenotypes secondary to deletions of 17p13.3. Am J Hum Genet. 2003;72(4):918-930. doi:10.1086/374320. PMID: 12621583. PMC1180354.
55. Boycott KM, Flavelle S, Bureau A, Glass HC, Fujiwara TM, Wirrell E, et al. Homozygous deletion of the very low density lipoprotein receptor gene causes autosomal recessive cerebellar hypoplasia with cerebral gyral simplification. Am J Hum Genet. 2005;77(3):477-483. doi:10.1086/444400. PMID: 16080122. PMC1226212.
56. Keays DA, Tian G, Poirier K, Huang GJ, Siebold C, Cleak J, et al. Mutations in alpha-tubulin cause abnormal neuronal migration in mice and lissencephaly in humans. Cell. 2007;128(1):45-57. doi:10.1016/j.cell.2006.12.017. PMID: 17218254. PMC1885944.
57. Corrigendum to: Definitions and classification of malformations of cortical development: practical guidelines. Brain. 2020;143(10):e108. doi:10.1093/brain/awaa298. PMID: 32968766.
58. Di Donato N, Chiari S, Mirzaa GM, Aldinger K, Parrini E, Olds C, et al. Lissencephaly: Expanded imaging and clinical classification. Am J Med Genet A. 2017;173(6):1473-1488. doi:10.1002/ajmg.a.38245. PMID: 28440899. PMC5526446.
59. Ciríaco MFM, Camara GF, Ribeiro EM. Síndrome de Miller-Dieker. In: Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 42–4.
60. Irahara K, Saito Y, Sugai K, Nakagawa E, Saito T, Komaki H, et al. Pontine malformation, undecussated pyramidal tracts, and regional polymicrogyria: a new syndrome. Pediatr Neurol. 2014;50(4):384-388. doi:10.1016/j.pediatrneurol.2013.12.013. PMID: 24507697. PMC3959267.
61. Kodani A, Kenny C, Lai A, Gonzalez DM, Stronge E, Sejourne GM, et al. Posterior Neocortex-Specific Regulation of Neuronal Migration by CEP85L Identifies Maternal Centriole-Dependent Activation of CDK5. Neuron. 2020;106(2):246-255.e6. doi:10.1016/j.neuron.2020.01.030. PMID: 32097629. PMC7255387.
62. Saillour Y, Carion N, Quelin C, Leger PL, Boddaert N, Elie C, et al. LIS1-related isolated lissencephaly: spectrum of mutations and relationships with malformation severity. Arch Neurol. 2009;66(8):1007-15. doi:10.1001/archneurol.2009.149. PMID: 19667223.
63. National Center for Biotechnology Information. ClinVar [Internet]. Bethesda (MD): NCBI; VCV000008074, VCV000160146, VCV000007071 [cited 2026 Oct 9]; MACF1: VCV000586948, VCV000586949, VCV000586950, VCV000586951, VCV000586953, VCV002499569, VCV003898866, VCV005020684; CEP85L: VCV000872889, VCV000872890, VCV000872891, VCV000872892, VCV000872893 [cited 2026 Nov 7]. Available from: https://www.ncbi.nlm.nih.gov/clinvar/
64. Mahendran G, Brown JA. Understanding the Molecular Basis of Miller-Dieker Syndrome. Int J Mol Sci. 2025;26(15):7375. doi:10.3390/ijms26157375. PMID: 40806509. PMC12347932.
65. Currier SA, Relland LM, Tobias JD. Perioperative Care of a Child With Miller-Dieker Syndrome. J Med Cases. 2026;17(7):297-302. doi:10.14740/jmc5320. PMID: 42327862. PMC13278722.
66. Baker EK, Brewer CJ, Ferreira L, Schapiro M, Tenney J, Wied HM, et al. Further expansion and confirmation of phenotype in rare loss of YWHAE gene distinct from Miller-Dieker syndrome. Am J Med Genet A. 2023;191(2):526-539. doi:10.1002/ajmg.a.63057. PMID: 36433683. PMC10099970.
67. Perrone M, Porro M, Piemontese P, Roggero P, Gangi S, Fumagalli M. Managing Drug-Resistant Epilepsy and Severe Feeding Difficulties in Miller-Dieker Syndrome: The Role of Enteral Ketogenic Diet. Case Rep Pediatr. 2026;2026:8660691. doi:10.1155/crpe/8660691. PMID: 42763564. PMC13589161.
68. Firth HV, Hurst JA. Deletions and duplications (including microdeletions and microduplications). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 661–3.
69. Bahi-Buisson N, Souville I, Fourniol FJ, et al. New insights into genotype-phenotype correlations for the doublecortin-related lissencephaly spectrum. Brain. 2013;136(Pt 1):223-44. doi:10.1093/brain/aws323. PMID: 23365099. PMC3562079.
70. Leger PL, Souville I, Boddaert N, et al. The location of DCX mutations predicts malformation severity in X-linked lissencephaly. Neurogenetics. 2008;9(4):277-85. doi:10.1007/s10048-008-0141-5. PMID: 18685874.
71. Kato M, Das S, Petras K, et al. Mutations of ARX are associated with striking pleiotropy and consistent genotype-phenotype correlation. Hum Mutat. 2004;23(2):147-59. doi:10.1002/humu.10310. PMID: 14722918.
72. Zhang C, Liang D, Ercan-Sencicek AG, Bulut AS, Cortes J, Cheng IQ, et al. Dysregulation of mTOR signalling is a converging mechanism in lissencephaly. Nature. 2025;638(8049):172-181. doi:10.1038/s41586-024-08341-9. PMID: 39743596. PMC11798849.
73. Huang R, Fu F, Zhang N, Zhou H, Mei S, Han J, et al. Genetic landscape and phenotypic correlations of lissencephaly: prenatal and postnatal insights. Brain Commun. 2026;8(2):fcag069. doi:10.1093/braincomms/fcag069. PMID: 41853045. PMC12993814.
74. Zillich L, Gasparotto M, Rossetti AC, Fechtner O, Maillard C, Hoffrichter A, et al. Capturing disease severity in LIS1-lissencephaly reveals proteostasis dysregulation in patient-derived forebrain organoids. Nat Commun. 2025;16(1):9091. doi:10.1038/s41467-025-64980-0. PMID: 41083500. PMC12518603.
75. Bölsterli BK, Steindl K, Kottke R, Steinfeld R, Boltshauser E. Lissencephaly with Brainstem Hypoplasia and Dysplasia: Think MACF1. Neuropediatrics. 2021;52(3):227. doi:10.1055/s-0040-1722690. PMID: 33511591.
76. Lei XY, Zhang MW, Sun H, Song W, Liang XY, Wang CS, et al. Identification of MACF1 as a causative gene of generalised epilepsy. J Med Genet. 2025;62(7):484-493. doi:10.1136/jmg-2025-110699. PMID: 40350249. PMC12322448.
77. Xi J, Deng F, Liang M, Ding Y, Li X, Gu Z, et al. Global developmental delay and focal seizures in individuals with de novo truncating MACF1 variants. Hum Genomics. 2026;20(1):38. doi:10.1186/s40246-026-00917-y. PMID: 41629993. PMC12918444.
78. Contrò G, Micalizzi A, Giangiobbe S, Caraffi SG, Zuntini R, Rosato S, et al. Posterior Lissencephaly Associated with Subcortical Band Heterotopia Due to a Variation in the CEP85L Gene: A Case Report and Refining of the Phenotypic Spectrum. Genes (Basel). 2021;12(8):1208. doi:10.3390/genes12081208. PMID: 34440382. PMC8391275.
79. Leduc-Pessah H, White-Brown A, Miller E, McMillan HJ, Boycott KM. Further characterization of CEP85L-associated lissencephaly type 10: Report of a three-generation family and review of the literature. Am J Med Genet A. 2023;191(12):2878-2883. doi:10.1002/ajmg.a.63380. PMID: 37621218.
80. Mattiske T, Moey C, Vissers LE, Thorne N, Georgeson P, Bakshi M, et al. An emerging female phenotype with loss-of-function mutations in the aristaless-related homeodomain transcription factor ARX. Hum Mutat. 2017;38(5):548-55. doi:10.1002/humu.23190. PMID: 28150386.
81. Proepper CR, Schuetz SM, Schwarz LM, Au KV, Bast T, Beaud N, et al. Characterization of the Epileptogenic Phenotype and Response to Antiseizure Medications in Lissencephaly Patients. Neuropediatrics. 2024;55(6):410-419. doi:10.1055/s-0044-1789014. PMID: 39214127.
82. Kolbjer S, Martin DA, Pettersson M, Dahlin M, Anderlid BM. Lissencephaly in an epilepsy cohort: Molecular, radiological and clinical aspects. Eur J Paediatr Neurol. 2021;30:71-81. doi:10.1016/j.ejpn.2020.12.011. PMID: 33453472.
83. Firth HV, Hurst JA. Imaging in prenatal diagnosis. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 742–3.
84. Firth HV, Hurst JA. Neonatal encephalopathy and intractable seizures. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 246–51.
85. Firth HV, Hurst JA. Epilepsy in infants and children. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 410–5.
86. Firth HV, Hurst JA. Intellectual disability—X-linked inheritance. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 194–7.
87. Firth HV, Hurst JA. Seizures with developmental delay/intellectual disability. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 308–10.
88. Firth HV, Hurst JA. Cerebellar anomalies. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 92–5.
89. Gamber A, Garton H, Parmar H, Armour E, Benedetti GM. Pearls & Oy-sters: Acute Obstructive Hydrocephalus Co-Occurrence With Vigabatrin-Associated MRI Changes in an Infant With LIS1-Related Classical Lissencephaly. Neurology. 2025;105(8):e214181. doi:10.1212/wnl.0000000000214181. PMID: 41004695.
90. Simoes C, Graña M, Rodriguez S, Baltar Yanes F, Tapié A, Dell'Oca N, et al. Novel frameshift mutation in LIS1 gene is a probable cause of lissencephaly: a case report. BMC Pediatr. 2022;22(1):545. doi:10.1186/s12887-022-03595-6. PMID: 36100855. PMC9472359.
91. Jones KL, Jones MC, del Campo M. Baraitser-Winter syndrome. In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 314–21.
92. Denommé-Pichon AS, Collins SC, Bruel AL, Mikhaleva A, Wagner C, Vancollie VE, et al. YWHAE loss of function causes a rare neurodevelopmental disease with brain abnormalities in human and mouse. Genet Med. 2023;25(7):100835. doi:10.1016/j.gim.2023.100835. PMID: 36999555.
93. Firth HV, Hurst JA. Structural intracranial anomalies. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 320–3.
94. Firth HV, Hurst JA. Ventriculomegaly. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 802–4.
95. Jones KL, Jones MC, del Campo M. Walker-Warburg syndrome. In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 256–7.
96. 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, section 1.5. doi:10.1007/978-3-662-63123-2_1.
97. 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. PMID: 22689735. PMC3369510.
98. 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. PMID: 35503712.
99. Mallinckrodt ARD LLC. Acthar Gel (repository corticotropin injection). Prescribing information [Internet]. Silver Spring (MD): U.S. Food and Drug Administration; revised 2026 Jul [cited 2026 Oct 9]. NDA 008372. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=7b48ddec-e815-45f4-9ca0-5c0daaf56f30
100. Lundbeck Pharmaceuticals LLC. SABRIL (vigabatrin) tablets and powder for oral solution. Prescribing information [Internet]. Silver Spring (MD): U.S. Food and Drug Administration; revised 2021 Oct [cited 2026 Oct 9]. NDA 020427 and NDA 022006 (original approvals 2009 Aug 21). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=a88ac1b4-e2c9-45c0-b321-4785902172e3
101. VIGAFYDE (vigabatrin) oral solution 100 mg/mL. Prescribing information [Internet]. Silver Spring (MD): U.S. Food and Drug Administration; revised 2025 Nov [cited 2026 Oct 9]. NDA 217684 (original approval 2024 Jun 17). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=8d3d6316-33ab-41e8-9485-4495c218be56
102. European Medicines Agency. Kigabeq (vigabatrin) 100 mg and 500 mg soluble tablets: EPAR — Product information (Summary of Product Characteristics) [Internet]. Amsterdam: EMA; marketing authorisation 2018 Sep 20 [latest EC decision 2026 Jul 22; cited 2026 Oct 9]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/kigabeq
103. AbbVie Inc. DEPAKOTE (divalproex sodium) delayed-release tablets. Prescribing information [Internet]. Silver Spring (MD): U.S. Food and Drug Administration; revised 2026 Mar [cited 2026 Oct 9]. NDA 018723. Available from: https://dailymed.nlm.nih.gov/
104. GlaxoSmithKline. LAMICTAL XR (lamotrigine) extended-release tablets. Prescribing information [Internet]. Silver Spring (MD): U.S. Food and Drug Administration; revised 2025 Oct [cited 2026 Oct 9]. NDA 022115. Available from: https://dailymed.nlm.nih.gov/
105. Bernasconi A, Martinez V, Rosa-Neto P, et al. Surgical resection for intractable epilepsy in "double cortex" syndrome yields inadequate results. Epilepsia. 2001;42(9):1124-9. doi:10.1046/j.1528-1157.2001.39900.x. PMID: 11580758.
106. Kurzbuch AR, Cooper B, Israni A, Ellenbogen JR. Non-pharmacological treatment options of drug-resistant epilepsy in subcortical band heterotopia: systematic review and illustrative case. Childs Nerv Syst. 2023;39(2):451-62. doi:10.1007/s00381-022-05638-w. PMID: 35933521.
107. Franco A, Pimentel J, Campos AR, Morgado C, Pinelo S, Ferreira AG, et al. Stimulation of the bilateral anterior nuclei of the thalamus in the treatment of refractory epilepsy: two cases of subcortical band heterotopia. Epileptic Disord. 2016;18(4):426-430. doi:10.1684/epd.2016.0878. PMID: 27965181.
108. Fukuoka M, Kuki I, Hattori Y, Tsuji H, Horino A, Nukui M, et al. Total callosotomy ameliorates epileptic activity and improves cognitive function in a patient with Miller-Dieker syndrome. Epilepsy Behav Rep. 2024;26:100670. doi:10.1016/j.ebr.2024.100670. PMID: 38725538. PMC11079450.
109. Dema A, Baskaran A, Chang J, Nordli D 3rd, Warnke P. Complete seizure freedom following minimally invasive laser callosotomy in a patient with lissencephaly type 1: Challenging the traditional paradigm. Epileptic Disord. 2026. doi:10.1002/epd2.70307. PMID: 42329172.
110. ClinicalTrials.gov [Internet]. Bethesda (MD): National Library of Medicine (US). Search: lissencephaly OR Miller-Dieker OR subcortical band heterotopia OR double cortex [cited 2026 Oct 9]; MACF1; CEP85L [cited 2026 Nov 7]. NCT05185414; NCT00041600. Available from: https://clinicaltrials.gov/
111. Verrillo L, Iannotti FA, Drongitis D, Martinello K, Poeta L, Barra A, et al. Cannabidiol attenuates epileptic phenotype and increases survival in a mouse model of developmental and epileptic encephalopathy type 1. Epilepsia. 2025;66(10):4035-4052. doi:10.1111/epi.18522. PMID: 40608247. PMC12605674.
112. Nagata T, Shibata T, Tsuchiya H, Akiyama M, Kato M, Akiyama T, et al. Immediate Therapeutic Response to Vigabatrin in Lissencephaly-Related Epileptic Spasms due to TUBA1A R402H Variant. Am J Med Genet A. 2025;197(7):e64030. doi:10.1002/ajmg.a.64030. PMID: 39996381.
113. Hou PS, Lim SJ, Lu HE, Tsai MH. Generation of an induced pluripotent stem cell line (KCGMHi001-A) from a patient with CEP85L related posterior predominant lissencephaly (LIS10). Stem Cell Res. 2026;94:104004. doi:10.1016/j.scr.2026.104004. PMID: 42066409.
114. Falsaperla R, Marino SD, Marino S, Pavone P. Electroclinical Pattern and Epilepsy Evolution in an Infant with Miller-Dieker Syndrome. J Pediatr Neurosci. 2018;13(3):302-307. doi:10.4103/jpn.jpn_182_17. PMID: 30271461. PMC6144613.
115. Gleeson JG, Minnerath S, Kuzniecky RI, et al. Somatic and germline mosaic mutations in the doublecortin gene are associated with variable phenotypes. Am J Hum Genet. 2000;67(3):574-81. doi:10.1086/303043. PMID: 10915612. PMC1287517.
116. Moreira I, Bastos-Ferreira R, Silva J, Ribeiro C, Alonso I, Chaves J. Paternal transmission of subcortical band heterotopia through DCX somatic mosaicism. Seizure. 2015;25:62-64. doi:10.1016/j.seizure.2014.12.005. PMID: 25645638.
117. Moloney PB, Benson KA, Phelan E, O'Regan M, Redmond J. Familial posterior predominant subcortical band heterotopia caused by a CEP85L missense mutation. Seizure. 2022;103:58-60. doi:10.1016/j.seizure.2022.10.018. PMID: 36306705.
118. Torres FR, Montenegro MA, Marques-De-Faria AP, Guerreiro MM, Cendes F, Lopes-Cendes I. Mutation screening in a cohort of patients with lissencephaly and subcortical band heterotopia. Neurology. 2004;62(5):799-802. doi:10.1212/01.wnl.0000113725.46254.fd. PMID: 15007136.
119. LISS e.V. — Hilfeforum Lissenzephalie [Internet]. Germany; [cited 2026 Oct 9]. Available from: https://lissenzephalie.de/
120. MedlinePlus Genetics. Miller-Dieker syndrome [Internet]. Bethesda (MD): National Library of Medicine (US); [cited 2026 Oct 9]. Available from: https://medlineplus.gov/genetics/condition/miller-dieker-syndrome/
121. de Wit MC, de Rijk-van Andel J, Halley DJ, Poddighe PJ, Arts WF, de Coo IF, et al. Long-term follow-up of type 1 lissencephaly: survival is related to neuroimaging abnormalities. Dev Med Child Neurol. 2011;53(5):417-21. doi:10.1111/j.1469-8749.2011.03937.x. PMID: 21410694.