Fetal malformations of cortical development: review and clinical guidance

Abstract

Malformations of cortical development (MCDs) are a heterogeneous family of congenital brain malformations that originate from disturbed development of the cerebral cortex. MCDs can arise from primary genetic disorders that lead to dysfunction of the molecular processes controlling neuronal proliferation, neuronal migration, cortical folding or cortical organization. MCDs can also result from secondary, disruptive causes, such as congenital infection or other in utero brain injuries. Sequelae of MCDs can include epilepsy, intellectual disability and cerebral palsy, among other symptoms, with a high burden of paediatric morbidity. Advances in antenatal genetic testing and imaging have improved the ability to diagnose MCDs, yet limited literature exists to aid clinicians in prognostication of outcomes and perinatal management. These clinical realities can make it challenging for clinicians caring for fetal neurological conditions to counsel families and make recommendations for interdisciplinary care. We aim to review the literature on fetal MCDs and present practice guidelines for clinicians regarding the pre- and postnatal management of MCDs.

As antenatal screening improves, an increasing number of congenital brain malformations can be identified prenatally. Russ et al. review the causes and consequences of fetal malformations of cortical development, and present clinical practice guidelines for their pre- and postnatal management.

Introduction

As antenatal screening via genetic testing and fetal imaging continually improves, a greater number of congenital brain malformations can be identified prenatally. Malformations of cortical development (MCDs), a family of congenital brain anomalies specifically affecting the cerebral cortex and often the cerebrum more broadly, are among the most common prenatally detected fetal brain malformations,^1,2 although the exact incidence is not known. MCDs are associated with significant long-term morbidity, including neurodevelopmental impairment, intellectual disability, cerebral palsy and epilepsy.^3,4

The heterogeneity of MCD types and locations, the variability in diagnostic precision based on imaging techniques, the wide spectrum of developmental outcomes and our ever-expanding understanding of the aetiologies and pathogenesis of MCDs can prove challenging for counselling. Prenatal counselling must balance appropriate gravity while acknowledging uncertainty around the spectrum of potential outcomes, particularly given the paucity of literature on the prenatal evolution of MCDs across gestational ages and the limited availability of detailed outcomes data. The timing of prenatal detection of MCDs and the confidence in diagnosis to then provide an accurate prognosis can all influence patient preferences for pregnancy management, including decisions about pregnancy continuation, location of delivery and management of labour and delivery.⁵

Existing literature extensively describes typical cortical development⁶ and the pathogenesis of MCDs,^7-9 but our understanding of the prenatal detection rate of MCDs and their subsequent outcomes remains limited.³ We aim to integrate a discussion of this literature and present practice recommendations for clinicians involved in perinatal care to manage and counsel expectant parents whose fetus is prenatally identified to have an MCD.¹⁰ Research gaps and priorities for future research are also addressed to guide further studies to advance our understanding of MCDs and guide fetal neurological care.

Brief summary of cortical development

Understanding the process of cortical development is essential for understanding which features of MCDs can be detected at various stages of gestation, in turn informing decisions about the timing of imaging, genetic testing and other clinical decision-making. A timeline of typical human cortical development and the onset of MCDs is outlined in Fig. 1. Briefly, excitatory neurogenesis in the human cortex begins in the ventricular zone of the dorsal pallium around 6 to 8 gestational weeks and proceeds through several stages until around 24 weeks. Radial glial cells in the ventricular zone give rise to intermediate progenitors and early neuronal precursors in the subventricular zone, which in turn amplify and diversify cortical excitatory neurons. In humans and non-human primates, an additional progenitor population, termed outer radial glia, become the primary neurogenic precursor population by 17 gestational weeks⁸ and are thought to represent an evolutionarily recent neuronal population that enhances human cortical neuron diversity. Excitatory neuron precursors give rise to the cortical plate, which, after several rounds of neurogenesis and outward radial migration that peaks around 20–22 weeks, develops into a canonical six-layer structure.^1,8 Radial migration of cortical excitatory neurons is classically described as following an ‘inside-out’ pattern,¹ whereby neurons generated earlier reside in deeper layers of cortex, although recent lineage tracing studies suggest some subtype-specific variations in this pattern.¹¹

Figure 1.

Timeline of typical corticogenesis or malformations of cortical development. Timeline of events in typical corticogenesis (top) or MCD onset (bottom). GW = gestational weeks; MCDs = malformations of cortical development.

Cortical folding, or gyrification and sulcation, produces highly stereotyped primary sulci, such as the central sulcus and Sylvian fissure, followed by secondary gyri and sulci, and finally tertiary gyri and sulci that demonstrate more inter-individual variability.¹² Initial formation of primary sulci, namely the proto-Sylvian fissure, is subtly detectable on fetal postmortem studies as early as 9–12 gestational weeks,¹³ although there is a lag before their detectability on fetal MRI. For example, the Sylvian fissure and parieto-occipital sulcus become observable on fetal imaging as early as 18–20 weeks.^12,14,15 Subsequent gyrification continues to mature throughout the late second and third trimesters with ongoing cortical neurogenesis and neuronal migration and is thought to be driven in part by differences in inter-areal neurogenesis and neuronal density, in conjunction with the mechanical forces of tangential cortical expansion.^8,16

In parallel to neurogenesis and neuronal migration, axons of the corpus callosum, which are derived largely from cortical excitatory neurons that project within the telencephalon, begin to cross the midline to connect the cortical hemispheres between 12 and 20 weeks.⁸ In contrast to excitatory cortical neurons, inhibitory interneurons are generated in the subpallium, ventral to the pallium, in the medial and caudal ganglionic eminences, and tangentially migrate into the cortical plate from approximately 20 weeks into the first postnatal year.^8,17,18 As cortical neurogenesis and migration taper off, cortical organization via early synaptogenesis begins around 22 weeks,^6,8 and persists postnatally until about 2 years of age when synaptic density peaks, followed by further synaptic pruning into adolescence.⁸ Finally, gliogenesis of astrocytes and oligodendrocytes from radial glial cells begins around 23 gestational weeks, peaking around 28 weeks, and axon myelination by oligodendrocytes begins between 23–32 weeks and matures through adolescence.⁸

Frameworks for categorizing fetal malformations of cortical development

There are two primary axes upon which MCDs are often considered^3,7,8,19: (i) the neurodevelopmental stage at which the MCD originates; or (ii) whether the MCD is primary (genetic) or secondary (disruption). The stage that the MCD originates is a pivotal determinant of the type and extent of MCD. While the classification of MCDs as primary or secondary can be an oversimplification, this distinction is a clinically relevant approach to guide prenatal diagnostic work-up and counselling. Still, it is important to recognize that similar convergent phenotypes can have primary and disruptive aetiologies, and there can be underlying genetic predispositions for disruptive brain injuries that give rise to MCDs.

Primary disorders

Classification of MCDs by neurodevelopmental stage is commonly organized into three major categories depending on the phase of cortical development most impacted: neuronal proliferation, migration or organization (sometimes referred to as post-migration development).^1,7,19 During cortical neuronal proliferation, wherein radial glial cells and intermediate progenitors expand the population of post-mitotic excitatory neurons, neuronal underproduction or enhanced apoptosis can result in microcephaly, whereas disorders such as hemimegalencephaly can result from neuronal overproduction or reduced apoptosis. Commonly disrupted molecular pathways in proliferating neuronal progenitors include those that regulate the formation and organization of the centrosomes and mitotic spindle (Table 1).⁸

Table 1.

Overview of malformations of cortical development by stage of cortical development

Cortical developmental process	MCD	Associated syndromes	Example gene families and molecular pathways	Example genes	Secondary causes
Excitatory neuronal proliferation	Microcephaly	–	Centrosome assembly and mitotic spindle formation	MCPH1, ASPM, NDE1, CENPJ	Exposure to teratogens
		–	Cell cycle/mitosis	MCPH1, NDE1, CENPJ, CDK5RAP2	Intrauterine infection
		–	Microtubule stability and transport	LIS1, DCX, TUBA1A, TUBB2B, TUBB3	–
		–	Transcription	FOXG1	–
	Megalencephaly	Sturge–Weber, Klippel–Trenaunay	Mitotic spindle formation	WDR62	–
		Neurofibromatosis 1, tuberous sclerosis	mTOR pathway regulation	PIK3R2, PIK3CA, AKT3	–
Excitatory neuronal migration	Lissencephaly	–	Microtubule stability and transport	LIS1, DCX, TUBA1A, TUBB2B, TUBB3, DYNC1H1	–
		–	Actin regulation	ACTG1	–
		–	Interneuron proliferation and migration	ARX	–
	Cobblestone malformation	Congenital muscular dystrophies	Glycosylation of alpha-dystroglycan and basement membrane integrity	LGMD2I, LGMD2K, LGMD2M, LAMB1, LAMB2, LAMC3	–
		Walker Warburg syndrome	–	POMT1, POMT2, POMGNT1	–
		–	–	GPR56	–
	Heterotopias	–	Microtubule stability and transport	TUBG1, MAP1B, DCX	–
		–	Actin regulation	FLNA, ACTG1	–
		–	Vesicle trafficking and fusion	ARFGEF2	–
		–	Ventricular zone protocadherins	FAT4, DCHS1	–
Cortical organization	Focal cortical dysplasia	–	mTOR pathway regulation	MTOR, AMPK, AKT3, PIK3CA, PTEN, DEPDC5	–
		–	Interneuron proliferation and migration	ARX	–
	Tuberous sclerosis complex	Tuberous sclerosis	mTOR pathway regulation	TSC1, TSC2	–
	Polymicrogyria	Zellweger, peroxisomal disorders	Mitotic spindle formation	WDR62	Exposure to teratogens
		–	Microtubule stability and transport	TUBA1A, TUBB2B, TUBB3, DYNC1H1	Intrauterine infection
		–	mTOR pathway regulation	AKT3, PIK3CA, PIK3R2, PIK4A	Intrauterine vascular insult
Cortical interruption	Schizencephaly/ porencephaly	Aicardi syndrome, Adam–Oliver syndrome,	Collagen stability, vascular integrity	COL4A1, COL4A2, COLGALT1, CTSA, FOXC1, NOTCH3, TREX1	Intrauterine vascular insult

Example syndromes, molecular pathways and single-gene aetiologies are listed for primary genetic causes of malformations of cortical development (MCDs) (note: syndromes are not necessarily directly correlated with molecular pathways or genes listed in the adjacent columns). Common acquired aetiologies of select MCDs are listed in the final column.

Primary disorders of neuronal migration, which disrupt radial migration of excitatory precursors and/or tangential migration of inhibitory precursors, can result in abnormal gyrification, including lissencephaly (absent gyri or smooth cortex) and pachygyria (abnormally large/thick gyri), as well as some cases of polymicrogyria (abnormally small, misfolded gyri).^7,8 Migration disorders can also result in misplaced neurons that do not transit to their correct final location, resulting in heterotopic clusters of grey matter, including subcortical band heterotopias, periventricular nodular heterotopias or focal heterotopias. Over-migration of neurons beyond the pial surface results in cobblestone malformation.¹ Often, genetic disturbances in microtubule subunits or other regulators of the neuronal cytoskeleton result in disorders of neuronal migration (Table 1).⁸

Finally, disorders of cortical organization result from inadequate or inappropriate synaptogenesis and cortical microcircuit formation. Many focal cortical dysplasias (FCDs), for example, arise from immature or dysplastic neurons, often in conjunction with inappropriately incorporated inhibitory interneurons, which together lead to a focal, hyperexcitable microcircuit capable of epileptogenesis.⁸ Primary polymicrogyria has typically been considered a disorder of cortical organization, though it is increasingly thought to result from late disturbances in neuronal migration.¹

While these schemata provide a helpful foundation for organizing the aetiologies of diverse MCDs, some MCDs may not fit cleanly into these categories. For example, mTORopathies, a family of primary genetic disorders that result from disturbances of molecular pathways regulated by ‘mammalian Target of Rapamycin’ (mTOR), may result in a wide phenotypic spectrum, from megalencephaly, a proliferative disorder (as in the case of STRADA-related polyhydramnios, megalencephaly and symptomatic epilepsy), to tuberous sclerosis complex (TSC) or FCD, which are cortical organization disorders.⁸ Likewise, the spectrum of tubulinopathies, genetic disorders associated with microtubule formation that were previously thought to primarily affect neuronal migration, has expanded to include phenotypes associated with disruption of cortical organization, such as polymicrogyria or dysgyria and aberrant axonal pathfinding.²⁰

Secondary disorders

Secondary disruptions in cortical development can also result in MCDs. Prenatal cerebral infection, fetal ischaemic infarct or haemorrhage, as well as in utero exposure to teratogens, can result in secondary microcephaly, polymicrogyria, schizencephaly, porencephaly, ventriculomegaly and callosal dysgenesis.¹ Schizencephaly is characterized by grey matter-lined clefts spanning the cerebral mantle from pia to ventricle while porencephaly is not lined by grey matter, although these entities fall along a spectrum. Although the literature is limited, schizencephaly is hypothesized to result from a vascular insult, typically in the territory of the middle cerebral artery,²¹ prior to neuronal migration, allowing residual migrating neurons to reach the schizencephalic edge creating the grey matter lined cleft.^1,22,23 Porencephaly, on the other hand, is thought to result from a vascular or infectious destructive process after neuronal migration.^1,22,23 Though typically thought of as secondary disruptions, schizencephaly and porencephaly can also be associated with insults due to primary genetic or syndromic conditions, such as Aicardi syndrome, Adam–Oliver syndrome, Galloway–Mowat syndrome or pathogenic variants in COL4A1/2, among others.^21-23

Disruption-related polymicrogyria is most commonly a result of congenital cytomegalovirus (CMV) infection,²⁴ which infects all cortical cell types but appears to have enhanced tropism for radial glial cells and may therefore specifically impact excitatory neuron proliferation and migration.²⁵ Polymicrogyria can also arise from sequelae of other infections, and in utero insults such as twin-twin transfusion syndrome (TTTS), co-twin demise, intraparenchymal haemorrhage, hypoxic-ischaemic injury or arterial ischaemic stroke, among other causes.^24,26,27 Clues that polymicrogyria is the result of an infectious aetiology over a primary genetic cause include calcifications, retinal pathology or broader systemic pathology postnatally.²⁴ These findings do not conclusively rule out genetic aetiologies, however, since conditions such as COL4A1/2 and Aicardi syndrome can also present with calcifications and/or retinal pathology. Finally, polymicrogyria that is located frankly within cerebral arterial territories, along cortical borderzone vascular areas, or on the border of an ischaemic infarct would suggest an acquired vascular aetiology.²⁴

Challenges to conventional frameworks

Microcephaly, defined as a head circumference at least three standard deviations below the mean for gestational age, is a phenotype that spans across these frameworks.^7,19,28 Primary microcephaly from a genetic aetiology can result from cortical neuron underproduction with secondary stunting of brain and skull growth.^8,19,28 Postmigrational microcephaly is also a primary genetic disorder, but is organizational or neurodegenerative, wherein diminished head growth progresses over time into early childhood due to cumulative neuronal loss.¹⁹ Phenotypic and genetic overlap can be seen with congenital microcephaly and tubulinopathies (TUBA1A, TUBB2B, TUBB3, TUBG1).¹⁹ Microcephaly can also be secondary to perinatal brain injury in utero or in the neonatal period. Thus, microcephaly describes a heterogeneous convergent phenomenon of numerous genetic and secondary aetiologies and is not easily captured by a simple framework.¹⁹ Likewise, polymicrogyria can be due to COL4A1/2-related disorders or peroxisomal disorders and in these cases would be most accurately characterized as a disruption-related MCD resulting from a primary genetic cause.

Prenatal imaging

Sonography

Prenatal imaging findings must be contextualized within the developmental stage of cortical development. Most screening anatomy ultrasounds are obtained between 18 and 22 gestational weeks, during neurogenesis and radial migration of excitatory neurons, but prior to gliogenesis, callosal maturation, and the majority of cortical gyrification.^8,23 In some high-risk pregnancies, an early comprehensive anatomy ultrasound may be indicated as early as 11 gestational weeks.²⁹ Thus, the first detailed imaging assessment of fetal brain anatomy typically takes place before MCDs fully manifest. Clues toward a possible MCD on ultrasound may be non-specific and include abnormal head shape or size, abnormal ventricular shape or size, abnormal ganglionic eminences, abnormal gyrification for gestational age, suspected callosal anomalies or absence of the cavum septi pellucidi, a midline structure that becomes visible around 15–18 gestational weeks.^2,23,30 Irregular ventricular borders may suggest destructive periventricular lesions, disorders of cortical migration, such as periventricular nodular heterotopia, or disorders of cortical organization, such as the subependymal nodules of TSC.²³ Schizencephaly from a destructive event occurring prior to neuronal migration might be visible on ultrasound at this stage. The morphology of the Sylvian fissure and operculum may herald the first signs of an MCD. For example, a widened and shallow Sylvian fissure can be seen with lissencephaly and cobblestone malformation, while a deep Sylvian fissure is typically seen with peri-Sylvian polymicrogyria.^23,30 Disorders of cortical gyrification, particularly those that are secondary to infectious, inflammatory or vascular insults that occur in the late second or early third trimesters would not be detectable until later in gestation^2,23,30 after routine anatomy scans are complete. Dedicated neurosonography can provide greater resolution in cases that are concerning for possible MCDs or other brain malformations on screening ultrasound.³¹

Systemic findings that are common on ultrasound in fetuses with any brain malformation include fetal growth restriction, decreased fetal movements and polyhydramnios.³² Systemic findings may also yield clues to an underlying cause, such as hepatosplenomegaly or echogenic bowel in congenital infection,³³ hemihypertrophy in overgrowth disorders³⁴ or cardiac rhabdomyoma in TSC.

Fetal brain MRI

In clinical practice, fetal MRI is usually prompted by an abnormality on prenatal ultrasound, particularly to clarify subtle findings concerning for brain malformations including MCDs. Generally, fetal brain MRI is thought to be of higher yield after 22 gestational weeks when brain development and cortical gyrification is swiftly underway, although many MCDs are challenging to diagnose prenatally.^35,36 Prior to more robust cortical gyrification, the cortical plate may demonstrate abnormal increased or decreased sulcation, asymmetry to the contralateral hemisphere or thinning or blurring of the underlying subplate or intermediate zone, which may herald early signs of an MCD.³⁷ Fetal MRI can increase the yield of accurately detecting cerebral malformations over ultrasound from 7% to 25%.^38-42 In one study, fetal MRI detected around 85% of MCDs later confirmed on postnatal MRI with a specificity approaching 100% when the abnormality is seen in two planes.⁴³ Fetal MRI has similarly been shown to have 84% agreement with post-mortem histopathology for cerebral malformations broadly, the gold-standard confirmation for most structural brain malformations.⁴⁴ Discrepancies between fetal MRI and post-mortem histopathology are most common for callosal and posterior fossa abnormalities, largely related to tissue collection and handling prior to histological analysis rather than inaccuracy of diagnosis.⁴⁴ Indeed, a more in-depth evaluation of diagnostic errors on fetal MRI for any indication demonstrates an overall 7% error rate, although in 95%, the ultrasound report was also incorrect.⁴⁵ Errors on fetal MRI were more likely to under-diagnose than over-diagnose, and the reasons were wide-ranging, including variations in radiologist experience, subtlety or complexity of cases and image quality.⁴⁵ Although existing studies provide an overview of the added diagnostic advantage of fetal MRI for cerebral malformations, most were not specifically designed to evaluate the accuracy of detecting MCDs.

Because of subtle morphological features of MCDs on fetal MRI and their evolution during cortical development in utero, postnatally derived criteria appear to be of limited use for prenatal diagnosis. The utility of developing a prenatal imaging classification scheme is highlighted by a multicentre retrospective cohort describing cortical formation abnormalities in 356 fetuses.⁴⁶ One hundred and seventy-two (48%) had MR between 17–23 gestational weeks and 184/356 (52%) at ≥24 gestational weeks. Through expert panel review, the type of abnormality on fetal MRI was categorized as bilateral and symmetric in 164 (46%), bilateral and asymmetric in 65 (18%) and unilateral in 127 (36%).⁴⁶ Ventriculomegaly was also present in 86 (24%) fetuses and additional structural brain abnormalities were present in 196 (55%), which most frequently included corpus callosum abnormalities in 127 (36%) and cerebellar malformations in 60 (17%).⁴⁶ Repeat in utero MRI was available for 64 fetuses from the original cohort.⁴⁷ Among fetuses with two MRI exams, the classification of the cortical folding abnormality remained the same in 62% and changed in 38%.⁴⁷ The criteria used for classifying fetal cortical folding abnormalities in this study are proposed for use in future research studies that aim to link longitudinal fetal and postnatal imaging. In addition to laterality and symmetry, the proposed criteria include the following descriptions: large area of dysmorphic brain, excessive sulcation, reduced sulcation/gyration, trans-mantle cleft, focal distortion (focal bite, wart-like, saw-tooth) and enlarged hemisphere (focal, entire).⁴⁶

For additional guidance on neuroradiologic diagnosis of MCDs, we also recommend a recent review from an expert panel that discusses prenatal neuroimaging of brain malformations, including neuronal migration and organization disorders and post-migrational abnormalities.⁴⁸ A review from Lerman-Sagie et al.²³ comprehensively discusses prenatal neurosonography for the identification of congenital brain malformations, including MCDs. Here, we focus on the available literature on prenatal MCDs beyond case reports.

Ganglionic eminence anomalies

The ganglionic eminences (GEs) are a portion of the germinal matrix/germinal zone located between the thalamus and caudate. Though not part of the cortex, the GEs are important transient structures ventral to the cortex, with proliferative zones that produce a range of cortical interneurons that migrate into the developing cortex. Anomalies of the GEs can include enlargement (Fig. 2A and E) and cavitation and can be a harbinger of fetal MCDs, although there is currently little knowledge on how prenatally detected GE abnormalities may predict MCD.^48,49 A multicentre retrospective cohort of GE anomalies on fetal MRI in 60 cases described MCDs associated with GE anomalies at a mean of 23.1 gestational weeks (range 17–33 weeks).⁴⁹ Among fetuses with bilateral cavitations (n = 29), 7% had bilateral polymicrogyria, 86% had reduced gyrification and 79% had reduced opercularization. There were five cases with unilateral cavitations of GE on fetal MRI, with unilateral polymicrogyria in one case and reduced gyrification in another. Among cases with bilateral GE enlargement (n = 19), unilateral polymicrogyria was present in 5%, bilateral polymicrogyria in 10%, reduced gyrification in 47%, reduced opercularization in 52%, macrocephaly in 37% and microcephaly in 16%. Seven cases had unilateral GE enlargement, with hemimegalencephaly seen in 71%, unilateral polymicrogyria in 57%, and macrocephaly in 71%. Another case series reported 22 patients with abnormal GEs on fetal MRI (range 21–33 weeks), with additional inclusion criteria of abnormal fetal head size and extracranial findings on prenatal ultrasound.⁵⁰ There was evidence of MCD on fetal and/or postnatal MRI in 36% and haemorrhage in the GEs in one case. In a retrospective cohort of fetal cortical folding abnormalities,⁴⁶ abnormally large GEs with or without cavitation were present in 3/43 (7%) with bilateral, asymmetric areas of dysmorphic brain, 17/91 (19%) with bilateral symmetric reduced sulcation/gyrification and 1/43 (2%) with unilateral focal distortion. Larger studies with longitudinal imaging are needed to determine the predictive value of fetal GE anomalies for the presence and later development of MCD.

Figure 2.

Fetal MRI examples of malformations of cortical development. Hemimegalencephaly: axial (A) and coronal (E) T2-single shot fast spin echo (SSFSE) images from fetal MRI of a 24-week gestational age fetus with right hemimegalencephaly, demonstrating asymmetric enlargement of the right cerebral hemisphere with decreased cerebral mantle lamination, consistent abnormality of neuronal migration. There is asymmetric enlargement of the right ganglionic eminence (black arrow), consistent with abnormality of neuronal proliferation, and mild irregularity of the right cerebral cortex (white arrows), consistent with co-existing polymicrogyria. Schizencephaly: axial (B) and coronal (F) T2-SSFSE images of a 23-week gestational age fetus show a large left open-lipped schizencephalic cleft (black arrows) with a thin-walled associated ventricular diverticulum. Note the cleft is lined by dysplastic grey matter. Periventricular grey matter heterotopia: axial (C) and coronal (G) T2-SSFSE images of a 28-week gestational age fetus with severe ventriculomegaly secondary to aqueductal stenosis and nodularity of the ependymal margins (black arrows) consistent with extensive grey matter heterotopia. Polymicrogyria: axial (D) and coronal (H) T2-SSFSE images from a 27-week gestational age fetus with abnormal sulcation and diffuse irregularity of the right cerebral cortex (black arrows) consistent with polymicrogyria.

Polymicrogyria

Polymicrogyria on fetal MRI is challenging to identify and can manifest as increased or decreased gyrification depending on gestational age (Fig. 2A, D, E and H).⁴³ Radiologic description should include whether suspected polymicrogyria is unilateral or bilateral, focal or diffuse and the cortical location(s). In a study of fetal and postnatal brain MRI concordance, 11/13 cases (85%) of polymicrogyria on postnatal brain MRI were detected on fetal MRI at a median of 25 gestational weeks.⁴³ Polymicrogyria was bilateral and diffuse in 69%, bilateral and focal in 15% and unilateral and focal in 15%. Additional cerebral abnormalities were seen in all cases, most commonly abnormalities of the corpus callosum in 77%, ventriculomegaly in 54% and hindbrain abnormalities in 54%. Polymicrogyria was the only MCD on fetal MRI in 38% and associated with periventricular nodular heterotopia in 54% and/or schizencephaly in 23%. There was one case with encephalomalacia consistent with prior infarct and one case with diffuse parenchymal thinning with haemorrhage or calcification. Of 81 total cases reviewed, the sensitivity of fetal MRI for polymicrogyria was reported to be 85% [95% confidence interval (CI) 55%, 98%], with a specificity of 100%.⁴³

Schizencephaly

Schizencephaly appears on fetal neuroimaging as unilateral or bilateral transmantle clefts lined by dysplastic grey matter (distinguishing it from porencephaly which is not lined by grey matter) and are classified as open lipped or closed lipped (Fig. 2B and F). A retrospective case series reported 10 fetuses with schizencephaly on fetal brain MRI at a mean gestational age of 27.6 weeks.⁵¹ While the majority of clefts were open on fetal MRI, around half evolved to closed schizencephaly on postnatal MRI.⁵¹ Polymicrogyria was detected in 3/18 (17%) clefts on fetal MRI, while polymicrogyria separate from the cleft was seen in 11%.⁵¹ Additional findings included absent cavum septi pellucidi in 50% and fetal agenesis of the corpus callosum in one case with new postnatal diagnosis of bilateral periventricular heterotopia.⁵¹ More recently, a retrospective cohort from George et al.²¹ reported fetal MRI findings in 22 fetuses with schizencephaly. Open schizencephaly was present in 28/34 (82%) defects and occurred in the middle cerebral artery territory in two-thirds. Additional intracranial abnormalities were present in all fetuses, most commonly polymicrogyria in 59%, posterior fossa abnormalities in 55%, callosal abnormalities in half and parenchymal haemorrhage associated with the cleft in 41%.²¹

Heterotopia

Periventricular nodular heterotopia (PVNH) are detected along the margins of the lateral ventricles, are isointense to the germinal matrix, and can be indistinguishable from subependymal nodules in tuberous sclerosis on fetal imaging (Fig. 2C and G).¹⁵ Patterns of heterotopia include PVNH that can occur unilaterally or bilaterally, singly or multiply, or as subcortical band heterotopia with confluent heterotopic grey matter. Fetal MRI is reported to have lower sensitivity for PVNH compared to other MCDs.⁴³ Of 15 cases of PVNH identified on postnatal MRI in one case series, heterotopia were only detectable on fetal MRI in 10 of those subjects.⁴³ The sensitivity of fetal MRI for heterotopia was reported to be 73% and notably decreased to 44% when stratifying by gestational age <24 gestational weeks, likely due in part to the small size of heterotopia. PVNH was bilateral in two-thirds of cases on postnatal MRI. Among cases where heterotopia was missed on fetal MRI, two were unilateral and three were bilateral. Additional MRI abnormalities were seen in all cases, including corpus callosum abnormalities in 67%, ventriculomegaly in 60%, hindbrain abnormalities in 47% and encephalocele/meningocele in 13%. PVNH was the only MCD in 53%, was associated with polymicrogyria in 33% and co-occurred with both polymicrogyria and schizencephaly in 13%.

The limited sensitivity and specificity of fetal MRI for heterotopia has similarly been shown in retrospective cohorts of prenatally diagnosed myelomeningocele.^52,53 Flanders et al.⁵² found that 99/497 (20%) fetuses with myelomeningocele had comorbid PVNH on fetal MRI, although 35/99 (35%) had confirmed PVNH on postnatal MRI.⁵² Conversely, of the 398 patients for whom heterotopia was not seen on fetal MRI, 12% had heterotopia on postnatal MRI.⁵²

Pachygyria-lissencephaly

Abnormal decreased gyrification on fetal imaging raises concern for an MCD along the pachygyria-lissencephaly spectrum. However, lissencephaly is typically only diagnosed after 27–30 gestational weeks, through demonstration of absent or abnormal development of the primary sulci, which should be radiographically detectable by 18–20 weeks.^12,15 As an example, the case in Fig. 3 illustrates this, where fetal MRI at 22 gestational weeks shows agenesis of the corpus callosum with shallow Sylvian fissures, but lissencephaly and subcortical band heterotopia become detectable upon postnatal MRI. A prospective study of fetal MRI in pregnant patients who had a prior fetus or child with lissencephaly enrolled 19 participants spanning 23 pregnancies; lissencephaly was detected in three of these cases (13%) and associated with hypogenesis of the corpus callosum in one and germinolytic cysts in another.⁵⁴ Four cases (17%) were classified as having a minor sulcation delay on the initial fetal MRI at 20–24 weeks, with subsequently normal sulcation on repeat fetal MRI at 27–28 weeks and/or 30–34 weeks.⁵⁴

Figure 3.

Pre- and postnatal imaging from a patient with subcortical band heterotopia. Coronal T2-weighted (A) fetal MRI at 22 gestational weeks, showing agenesis of the corpus callosum and shallow Sylvian fissures with absence of the expected multi-layered pattern. Axial T1-weighted (B) postnatal MRI showing subcortical band heterotopia and lissencephaly-pachygyria due to pathogenic DCX variant.

Cobblestone malformation, previously called type 2 lissencephaly, is an undersulcated cerebral surface with an irregular or ‘pebbled’ appearance and thickened cortex; however, current standard fetal MR techniques do not allow for identification of this cortical irregularity.⁴⁶ Rather, global decreased gyration is usually seen.⁴⁶ Typically associated with α-dystroglycanopathies, cobblestone malformation may show other concurrent imaging signs, including ventriculomegaly, abnormal operculization, irregular cortical thickness, abnormal lamination, callosal agenesis or dysgenesis, an angular (‘Z-shaped’) brainstem beyond 16 gestational weeks, diffuse cerebellar hypoplasia/dysgenesis, peripheral cerebellar ‘cysts’ (more conspicuous postnatally), ocular globe dysmorphology and other abnormalities such as retinal detachment and/or cataracts.²³ Lissencephaly due to tubulinopathies can also co-occur with findings such as microcephaly, callosal dysgenesis, brainstem and cerebellar hypoplasia/dysgenesis, GE enlargement, polymicrogyria and dysmorphic, incompletely separated basal ganglia.^23,55 One of the most important fetal MRI clues for a tubulinopathy is asymmetric malformation involving the cerebrum, brainstem and/or cerebellum.

Considerations for aetiologic investigation

Genetic

There are currently no formal guidelines to support a systematic approach to prenatal genetic testing in fetal MCDs. Some experts have proposed workflows for genetic testing and diagnosis in neonates and children with MCDs.³² Initial screening via chromosomal microarray (CMA) or single-nucleotide polymorphism (SNP) array can sometimes identify causative pathogenic copy number variants (CNVs). Several common CNVs are frequently associated with MCDs, including 22q11.2 deletion (often associated with polymicrogyria) and 6q deletion.³² Wang et al.⁵⁶ reported a retrospective cohort of 32 fetuses with MCDs detected by prenatal US and/or fetal brain MRI. Pathogenic CNVs on SNP array were detected in 9%.⁵⁶ Of these, two-thirds were found to have a 17p13.3p13.2 microdeletion involving the LIS1 gene with phenotypes consistent with Miller–Dieker lissencephaly syndrome.⁵⁶ For some specific MCDs, such as PVNH, the diagnostic yield of microarray can be as high as 36%.⁵⁷

In rare scenarios the clinical suspicion is high enough for targeted testing, such as testing DCX and LIS1 (PAFAH1B1) in the setting of subcortical band heterotopia. Broad genetic testing, however, is typically of higher utility and there are emerging data on combinatorial genetic testing in fetal brain malformations, including MCDs.⁵⁸ Exome sequencing (ES) has been shown in some cohorts of MCDs to increase the diagnostic yield beyond what is found on CMA; in one cohort, microarray identified likely explanatory variants in 5/54 (9%), while ES identified an additional 14.⁵⁹ A recent systematic review and meta-analysis⁶⁰ examined the yield of prenatal ES for any CNS anomaly on prenatal ultrasound or fetal MRI. The study included 1583 cases from 30 studies, case series and case reports.⁶⁰ The overall identification of a causative pathogenic or likely pathogenic variant with prenatal ES was 32% (95% CI 27%–36%).⁶⁰ Among 92 cases of fetal MCD, the incremental yield of prenatal ES was 35% (95% CI 26%–44%), with variants of TUBA1A and TUBB most commonly detected in the context of lissencephaly or complex MCDs associated with other brain malformations.⁶⁰ Wang et al.⁵⁶ reported an incremental yield of ES of 19/32 (59%) in a cohort of fetal MCD. Chromosomal or molecular diagnoses were made in all seven cases of lissencephaly, 4/8 (50%) of pachygyria, 4/8 (50%) of polymicrogyria and both cases of non-isolated heterotopia.⁵⁶ A retrospective cohort of prenatally diagnosed PVNH reported FLNA variants in 6/11 cases (55%) with diffuse heterotopia in whom genetic testing was obtained.⁶¹ Among fetuses with schizencephaly, variants of uncertain significance were reported in 2/22 (9%) cases in one cohort,²¹ and there are additional case reports of COL4A1 mutation in fetal schizencephaly.⁶² In a case series of 22 fetuses with GE anomalies, abnormal head size and extracranial anomalies, 15/17 (88%) who underwent genetic testing had a molecular diagnosis, most frequently tubulinopathies and MTOR variants.⁵⁰ The yield of genetic testing for isolated GE anomalies, however, is not known. Finally, the combination of microarray and ES can together result in a diagnostic yield of almost 50% for microcephaly.⁶³

Genome sequencing (GS), which can be used to assess both CNVs and single gene disorders, is beginning to be explored for prenatally detected MCDs. For example, Liao et al.⁶⁴ report GS in a prospective cohort of 28 fetuses with abnormal Sylvian fissures on neurosonography, of which 54% had MCDs. Among 23 cases with bilateral Sylvian fissure abnormalities, 70% had diagnostic variants on GS, including all cases with lissencephaly and/or pachygyria.⁶⁴ By contrast, the diagnostic yield of GS was 0/5 in cases with unilateral Sylvian fissure abnormalities.⁶⁴ In another example, GS in 12 children with undiagnosed lissencephaly identified three structural variants that would not be detected via ES or CMA.⁶⁵ While GS holds promise for improving detection of the genetic aetiologies of prenatally identified MCDs, it is not readily available in many clinical settings.

Risk factors for brain injury and secondary MCDs

Monochorionic multiple gestation increases the risk for fetal ischaemic and haemorrhagic brain injury with consequences that can include MCDs. These risks further increase in the setting of complications, namely twin-twin transfusion syndrome (TTTS) and co-twin demise. A population-based case-control study of twins with brain malformations reported MCDs in 39/56 (70%) monochorionic twins, with polymicrogyria in 37/39 (95%), schizencephaly in 8/39 (21%) and/or PVNH in 15/39 (38%).⁶⁶ Monochorionic twin gestation was also present in several cases of fetal schizencephaly in one cohort,²¹ and twins are over-represented in a population-based cohort of polymicrogyria.⁶⁷ Twin gestation also increases risk for germinal matrix haemorrhage,⁶⁸ which can precede the development of PVNH from ependymal injury. Griffiths et al.⁶⁹ reported a retrospective cohort of monochorionic twins complicated by co-twin demise, with demise occurring after laser therapy for TTTS in 27/68 cases (40%) and demise occurring spontaneously in 41/68 (60%). Fetal MRI abnormalities were identified in 9/58 cases (16%), with 6/9 (67%) showing more than one abnormality that included polymicrogyria (n = 5), schizencephaly (n = 1), remote infarction or encephalomalacia (n = 5) and microcephaly (n = 3).⁶⁹ Furthermore, Gebb et al.⁷⁰ reported abnormal fetal MRI findings in 30% of fetuses with co-twin demise after laser for TTTS and found abnormal MRI was more common after recipient demise than after donor demise (50% versus 14%). Whether dual diagnoses of pathogenic genetic variants compounded by other co-occurring risk factors for acquired brain injury are present in some twins with MCDs is not known.

Fetal sex is also a risk factor for some MCDs, as males are over-represented among some reported cohorts of polymicrogyria, with one cohort of 328 cases reporting a 3:2 predominance of males.⁷¹ Males were 1.5-fold more frequent than females in a cohort of 19 fetuses with schizencephaly for whom fetal sex was known.²¹ Sex effects are also reported in cohorts of fetal intracranial haemorrhage, with the proportion of males ranging from 57%–77%.^72-74 Given there are higher proportions of males with perinatal arterial ischaemic stroke, as well as intracranial haemorrhage,⁷⁵ the apparent predominance of males with polymicrogyria and schizencephaly may reflect a similar trend of sex effects in brain injuries across gestational ages.

Most congenital infections such as CMV are occult; however, screening for infectious symptoms, infectious risk factors and travel exposures can help provide clues about aetiologies of MCDs. Congenital infections associated with polymicrogyria, schizencephaly and microcephaly include CMV, toxoplasmosis, Parvovirus B19, Zika virus and lymphocytic choriomeningitis virus (LCMV).⁷⁶ Gestational age at the time of infection is a key driver of the type and extent of MCDs. Associated neuroimaging abnormalities like calcifications and periventricular cysts may point toward an infectious cause. LCMV is an underappreciated cause of congenital brain malformations that is transmitted by the common house mouse and more likely to affect low-income rural and urban pregnant persons.^77-79 Healthcare disparities⁸⁰ that disproportionately affect patients from low-income backgrounds, such as young maternal age, insufficient prenatal care or increased exposure to infectious vectors, may enhance the risks of MCDs.^81-83

Clinicians should also screen for other potential risk factors for secondary MCDs.⁸⁴ A maternal history of deep venous thrombosis or recurrent miscarriage may suggest thrombophilia such as antiphospholipid syndrome, with an increased risk for perinatal stroke and secondary MCDs. The use of teratogenic medications, while broadly associated with malformations of the CNS,⁸⁵ has not been specifically linked to MCDs, though sparse literature proposes a link between maternal warfarin use and schizencephaly.⁸⁶ Fetal exposure to recreational substance use, particularly alcohol, has been associated with numerous structural brain malformations,⁸⁷ including schizencephaly⁸³ and polymicrogyria.⁸⁸

Placental pathology is increasingly recognized as an important element of the evaluation of perinatal brain disorders. Histopathologic examination of placental specimens can help confirm presumed infectious aetiologies, such as CMV or Zika.^89,90 Evidence of maternal or fetal vascular malperfusion, ischaemia or haemorrhage may provide clues toward in utero vascular events contributing to acquired MCDs, such as schizencephaly or porencephaly.^91,92

Summary of outcomes

Postnatal outcomes for children with MCDs vary widely by the type, location and underlying aetiology. There are very few studies that specifically link the aetiology and outcomes of prenatally identified MCDs by precise subtype. Selected studies are summarized in Supplementary Table 1, focusing on prenatally identified malformations where possible. Among the challenges that arise when MCDs are detected prenatally, is that the full extent of the malformation cannot be known until later in brain development, and the underlying cause often has not yet been identified at the time of prenatal neurological consultation. These factors contribute to layers of diagnostic, aetiologic and prognostic uncertainty. Given the limited literature on the natural history of prenatally identified MCDs, clinicians counselling expectant parents may cautiously extrapolate the literature on postnatally identified MCDs, recognizing the inherent limitations of this approach. Several general themes related to prognosis are presented here, which may be helpful in preparation for counselling.

Nearly all MCDs can serve as an epileptogenic focus and thus increase the risk of epilepsy. Seizure onset in the neonatal period can occur in up to a quarter of patients with MCDs, particularly those with polymicrogyria, lissencephaly and other disorders of abnormal sulcation.⁴ One retrospective cohort reported neonatal seizure outcomes among neonates with prenatally or postnatally identified congenital brain malformations.⁴ Among 74 neonates that underwent continuous video EEG in this study, neonatal-onset epilepsy occurred in 9/34 (26%) with disorders of cortical neuronal migration/organization, 6/38 (16%) with disorders of early prosencephalic development and 2/16 (13%) with complex total brain malformations.⁴ Of the nine with MCDs and neonatal-onset epilepsy, six had polymicrogyria and three had pachygyria-lissencephaly.⁴ Neonatal seizures were electrographic only in five (28%), of whom 3/5 had MCDs. Of those with congenital brain malformations and neonatal seizures, 72% required ≥2 anti-seizure medications and 39% died within the neonatal period,⁴ emphasizing the need to identify those at highest risk for seizures at the time of prenatal diagnosis for more accurate counselling.

Generally speaking, the developmental outcomes from isolated focal MCDs tend to have a lower risk of neurodevelopmental impairment than multifocal or diffuse MCDs. For example, focal cortical dysplasia and isolated heterotopia can present in adolescents with focal epilepsy but otherwise age-appropriate development, in contrast to lissencephaly-pachygyria, or hemimegalencephaly, which often present with profound intellectual disability, cerebral palsy, refractory epilepsy and/or vision or hearing impairment from infancy.³ Likewise, focal disruption-related MCDs, such as unilateral porencephaly from an in utero infarct, often have lower risk of neurodevelopmental impairment than diffuse secondary aetiologies, such as congenital infection, or those that arise from primary genetic, multisystem and syndromic aetiologies. Cortical disruption, particularly from schizencephaly or porencephaly,⁹³ can lead to contralateral hemiparesis and motor deficits, while bilateral involvement can lead to quadriparesis and more severe cognitive and language delays. Existing data suggest that earlier age at seizure onset and concurrent microcephaly or megalencephaly are associated with increased risk of neurodevelopmental impairment, as is the presence of other malformations in addition to MCDs.^3,71

While current literature is often limited to small case series of heterogeneous MCDs and longitudinal multicentre studies are imperative for more accurate data on outcomes for each subtype of MCD, we highlight here some of the existing data on outcomes for the main MCD categories to help guide clinical prognostication.

Polymicrogyria

There are no studies that focus on the long-term outcome of prenatally identified polymicrogyria. There are, however, some retrospective studies that report clinical phenotypes of patients with polymicrogyria. A regional population-based cohort identified an overall prevalence of polymicrogyria in children of 0.02% [2.31 (95% CI 1.90–2.80) per 10 000].⁶⁷ Of the 109 children with polymicrogyria in this cohort, the median age at diagnosis was 1.3 years (interquartile range 0.3–3.0 years), with 44% of children diagnosed by 1 year.⁶⁷ Sixteen children in the cohort (15%) were diagnosed by MRI in the prenatal or neonatal period, although the exact number with suspected or confirmed prenatal diagnosis is not reported.⁶⁷ The most common symptoms prompting diagnosis were seizures in 28% and developmental delay in 24%.⁶⁷ The presence of bilateral polymicrogyria and concurrent congenital malformations were associated with earlier symptom onset and diagnosis.⁶⁷ Seizures occurred in 66 (61%) children, with onset during the first year in 26 (24%).⁶⁷ The presenting seizure type was variable, with 10% of children presenting with seizures in the setting of fever.⁶⁷ Neurodevelopmental disorders were present in the vast majority of children (103/109, 94%), with intellectual disability in 58 (53%) and developmental delay or learning difficulties in 20 (19%).⁶⁷ One-third of the cohort had cerebral palsy, of whom 31/36 (86%) were non-ambulatory, and other motor disorders like apraxia, hypotonia and developmental coordination disorder were present in 31/109 children (28%).⁶⁷

A multicentre retrospective cohort reported by Leventer et al.⁷¹ described 328 individuals with polymicrogyria. Of 189 individuals with known age at time of presentation, 38% presented antenatally or neonatally, although the proportion with abnormal prenatal ultrasound was not reported.⁷¹ Dysmorphic features or congenital anomalies other than polymicrogyria were observed in 73/166 (44%), and the majority had more than one abnormality, most commonly facial dysmorphisms (n = 42), hand, foot or digital abnormalities (n = 16), arthrogryposis or talipes equinovarus (n = 10), skin abnormalities (n = 8), palatal abnormalities (n = 7) and congenital heart defects (n = 6).⁷¹ Global developmental delay was described in 117/168 (70%) and isolated language delay in 32/168 (19%).⁷¹ Of 144 patients with neuropsychological data available, intellectual disability was present in 95 (67%).⁷¹ Global developmental delay and intellectual disability were more common in generalized polymicrogyria, whereas isolated language delay was more common in peri-Sylvian polymicrogyria.⁷¹ Epilepsy data were available in 225 cases, with a reported rate of 78% and mean age of onset of 4.9 ± 6.7 years, with a range of 1 day to 34 years.⁷¹ Neonatal seizures occurred in 14/225 (6%) and seizure-onset occurred within the first year in 57 (43%).⁷¹ Generalized polymicrogyria had a significantly lower age at seizure onset (median age 8 months) than other forms.⁷¹

Schizencephaly

Available studies on outcomes of schizencephaly detected prenatally are small. Of the retrospective cohorts of prenatally identified schizencephaly in the literature,^21,51,94 outcomes are only reported in one.²¹ Twenty-two fetuses with schizencephaly on fetal MRI were reported by George et al.²¹ with 36% liveborn, one stillborn (5%) and termination of pregnancy reported in 59%. One liveborn infant died at 1 day of age.²¹ Shunted hydrocephalus occurred in 1/7 (14%), epilepsy in 4/7 (57%) and gastrostomy-tube feeding in 1/7 (14%).²¹ Neurodevelopmental outcomes included cerebral palsy in 4/5 (80%), speech delay in 3/5 (60%) and intellectual disability in 1/5 (20%).²¹ Cerebral palsy was present in all cases with a persistently open defect (open: 4/4 versus closed 0/1).²¹ Children that developed epilepsy all had cortical malformations remote in location from the schizencephalic defect (4/4 versus 0/3 without epilepsy).²¹

Heterotopia

One retrospective cohort of 32 fetuses with prenatally diagnosed periventricular nodular heterotopia (PVNH)⁶¹ reported termination of pregnancy in 24 cases (80%), with follow-up available in six cases at a range of 6 months to 5 years. Among these six cases of diffuse PVNH, neurodevelopment was normal in five (83%) and one case had developmental delay and epilepsy.⁶¹ Another retrospective cohort found children with prenatally diagnosed myelomeningocele with PVNH and those without heterotopia, performed in the average range on the Bayley Scales of Infant and Toddler Development, 3rd or 4th edition, at age 1 year.⁵² In the setting of isolated PVNH, epilepsy occurs in the majority, with focal impaired awareness seizures commonly emerging in the second decade or later. Among a cohort of 34 patients with FLNA-related heterotopia followed for a median of 19.7 years, 24 (71%) had epilepsy.⁹⁵ Most adults with isolated PVNH are reported to have normal intelligence, although there are some reports of dyslexia being more common than in the general population.⁹⁶ A case-control study of 10 children with PVNH and 10 matched controls indicated that children with PVNH performed worse than controls on a task related to reading fluency.⁹⁷ Children with PVNH also showed worse adaptive skills than controls.⁹⁷

In contrast to PVNH, cognitive outcome in subcortical band heterotopia is commonly impaired. Males with subcortical band heterotopia in a retrospective cohort⁹⁸ ranged in age from 1 month to 34 years. The majority developed epilepsy (28/30, 93%), with an age of onset at a median of 36 months (range 5 days to 16 years).⁹⁸ Language delay or impairment was reported in 13/30 (43%).⁹⁸ Intellectual disability was present in 60%, of whom half were classified in the severe to profound range.⁹⁸ Cognition was classified as normal in six (20%) and borderline in six (20%).⁹⁸ A cross-sectional study of cognitive functioning in seven children (six female) with subcortical band heterotopia reported that impairments were observed broadly across cognitive and neuropsychological domains.⁹⁹ Intellectual ability was generally within the ‘extremely low’ range (full-scale IQ 44–74; performance IQ 45–72; verbal IQ 57–80).⁹⁹ Severe impairments were reported in processing speed, working memory and arithmetic.⁹⁹

Pachygyria-lissencephaly

There are no studies of long-term outcomes in prenatally identified pachygyria-lissencephaly. In one prospective cohort of lissencephaly,¹⁰⁰ 11/24 (46%) children were alive at a mean follow-up of 14 years. All surviving children showed severe intellectual disability, refractory epilepsy and complete dependency for self-care.¹⁰⁰ A multicentre retrospective cohort of LIS1-associated lissencephaly reported outcomes in 22 individuals aged 8 months to 24 years.¹⁰¹ All patients developed medication-refractory epilepsy, with the onset of seizures within the first 6 months in 82%, most frequently infantile spasms.¹⁰¹ In another retrospective cohort of LIS1-associated lissencephaly, 63 individuals of median age 6 years (range 1–39 years) were reported.¹⁰² This cohort demonstrated severe developmental deficits in nearly all subjects (90%), with severe motor impairment including spastic quadriparesis in 60%, nearly no language development in all cases and autism-related behaviours and sleep disorders in 30%.¹⁰² Seizures occurred in 21/38 (55%) and were medically refractory in all cases.¹⁰²

Practical recommendations for paediatric neurologists

A survey of paediatric neurologists in the United States revealed that their top priority for the evolving field of fetal neurology is the development of clinical practice guidelines for prenatal counselling and postnatal management.¹⁰ Given the limitations in the literature detailed above, further systematic research on large registry-based cohorts of patients with prenatally diagnosed MCDs will be paramount for the development of evidence-based guidelines. In the meantime, we aim to provide practical clinical recommendations developed by an international working group of paediatric neurologists with expertise in fetal neurological consultation, perinatal management and neurodevelopmental follow-up. These recommendations were developed with multidisciplinary input from paediatric neuroradiologists with expertise in fetal MRI, as well as maternal-fetal medicine, and paediatric genetics specialists with expertise in prenatal diagnosis. The expert opinion of this group is provided below to guide paediatric neurologists and other clinicians in the care of suspected or confirmed fetal MCDs.

Antenatal perinatal management

Fetal imaging

Fetal brain MRI should be offered in all cases of suspected MCD or other abnormal brain findings on prenatal ultrasound, consistent with recommendations from the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG),¹⁰³ ideally after 22 weeks in a centre with protocols optimized for the fetal brain. However, an earlier fetal MRI may be necessary depending on regional regulations regarding pregnancy decisions. Fetal MRI should also be considered in cases at increased risk for MCDs, such as monochorionic co-twin demise or TTTS treated with laser therapy. Though the optimal timing of imaging in this setting is not established, obtaining brain MRI at least 2 weeks post-procedure or co-twin demise might improve visualization of evolving injury and/or malformation.^27,104 Follow-up fetal MRI for better characterization of suspected MCDs should be considered on a case-by-case basis as available, particularly if the results would change pregnancy or perinatal management. Repeat fetal MRI may also be helpful if the initial findings are inconclusive; for instance, if abnormal Sylvian fissures, the early appearance of an abnormal sulcus or delayed sulcation is identified. Interpretation of fetal MRI by a neuroradiologist with expertise in fetal imaging is essential.

Pregnancy and family history

Maternal medication use (i.e. Warfarin), substance use (i.e. alcohol), risk factors for infection (i.e. travel to Zika-endemic region), occupational exposures (i.e. daycares or veterinary facilities), risk factors for perinatal stroke (i.e. history of thrombophilia) and history of trauma should be elicited. Obstetric history, including recurrent miscarriages, should also be discussed. A thorough family history of genetic, developmental and neurologic disorders, including epilepsy, should be obtained. History of consanguinity should also be elicited.

Aetiologic testing

Genetic

Carrier screening and aneuploidy screening (most commonly cell-free DNA) as available should be offered to all prenatal patients.¹⁰⁵ For congenital anatomic malformations or abnormal results of prenatal genetic screening, amniocentesis should be offered for CMA with or without a karyotype on fetal amniocytes. Based on availability and insurance coverage, a comprehensive gene panel that covers common monogenic causes of MCDs (Table 1), ES or GS should be offered when CMA is non-diagnostic, unless a secondary aetiology such as congenital infection or TTTS has been established. In cases of termination of pregnancy, genetic testing can be performed on the products of conception.

Infectious

Serology should be collected for toxoplasmosis, CMV (including avidity if CMV IgG is positive),¹⁰⁶ Zika virus, LCMV or other congenital infections, depending on history, risk factors and imaging findings. PCR testing for CMV and other congenital infections based on the history should be performed in cases where amniocentesis is performed.

Fetal neurology consultation

An interdisciplinary approach to care for patients with prenatally diagnosed MCDs is recommended.²³ Input from maternal-fetal medicine specialists, fetal/paediatric neurologists, radiologists who specialize in fetal neuroimaging, geneticists, neurosurgeons, palliative care clinicians and neonatologists can help provide information and consensus around diagnosis, aetiology and prognosis and additional planning for perinatal care for fetuses with MCDs. Fetal neurology consultation should synthesize the available diagnostic and prognostic information, as well as acknowledge uncertainty, the weak evidence for long-term outcomes and the substantial heterogeneity between individuals. Paediatric and perinatal palliative care clinicians can offer support around communication, decision-making and prognostic uncertainty. Palliative care consultation should be considered in all cases of MCDs and offered uniformly when available in cases where expectant parents face serious decision-making and/or significant prognostic uncertainty.¹⁰⁷

Timing, mode and location of delivery

Mode and timing of delivery should be individualized based on obstetric considerations. Location of delivery should be determined by likelihood of requirement for resuscitation, neonatal intensive care and multidisciplinary specialty care needs after delivery.¹⁰⁸ Delivery at a tertiary care centre is recommended for all prenatally identified MCDs, especially for cases with diffuse or more extensive MCDs, complex total brain malformation that includes MCDs or other systemic anomalies such as congenital heart disease. Otherwise, delivery at a tertiary care centre may not be necessary if expectant parents would prefer to perform postnatal evaluation, including MRI as an outpatient and if transport to a tertiary care centre after delivery is still feasible if emergently needed. Placental specimens should be preserved and sent for histopathology, particularly in cases where infectious or vascular placental pathology may corroborate potential secondary MCD aetiologies.

Postnatal management

Neonatal examination

Physical exam of the neonate for additional dysmorphic features and congenital abnormalities of other systems should be undertaken. Head circumference should be measured and trended over time. A systematic neurological exam should be performed. Baseline assessments by occupational/speech therapy and physical therapy should be performed as needed.

Neuroimaging

Postnatal imaging may include head ultrasound to establish baseline ventricle size or to exclude intraventricular haemorrhage if there is clinical concern.¹⁰⁹ Ultrasound may also identify abnormal gyration, midline defects, such as absence of the septum pellucidum, or calcifications.¹ A high-resolution brain MRI is the study of choice in the evaluation of MCD and should be obtained in an imaging facility with MRI protocols optimized for the neonatal and infant brain tailored to assess brain malformations. MRI (3 T preferred if available) ideally should be obtained in the neonatal period or within the first 3 months as a ‘feed and swaddle’ scan to avoid the requirement for anaesthesia or sedation.¹¹⁰ Follow-up brain MRI after 2 years of age, once myelination is near its mature appearance on imaging, may be considered if clinically warranted to evaluate the full extent of MCD. CT of the brain may be considered in cases in which there is a contraindication to MRI or MRI is not available, although the contrast resolution of MRI is far superior.

EEG

Continuous video EEG should be considered in neonates with prenatally identified MCDs for baseline assessment and seizure risk stratification, particularly when high-risk MCDs, such as polymicrogyria or lissencephaly, are present. EEG should be obtained in all neonates with encephalopathy or decreased arousal and in those with clinically suspected seizures.

Other investigations

Algorithms for postnatal genetic testing by clinical and radiologic phenotype have previously been published.³² Genetic testing with CMA and ES or GS should be obtained if prenatal genetic work-up was not performed. If broad sequencing is not available, targeted gene panels should be obtained as able. In cases with no established aetiology, or cases with inconclusive maternal serology (i.e. CMV IgG positive without avidity), testing to exclude CMV should be obtained within 3 weeks of birth. Some MCDs and patterns should prompt screening evaluation of other systems (Table 2). For instance, hypopituitarism and optic nerve hypoplasia should be ruled out in patients with MCDs where there is also evidence of midline anomalies or syndromes associated with septo-optic dysplasia. Baseline evaluation by ophthalmology for MCDs that commonly co-present with ocular and/or retinal anomalies should be performed. Infants and children should be referred to audiology for signs of hearing impairment and/or speech delay.

Table 2.

Specific considerations for selected malformations of cortical development and patterns

Malformation or pattern	Additional investigations and considerations
Selected MCDs
Megalencephaly/Hemimegalencephaly	Dermatologic examination for neurocutaneous features; Examination for signs of overgrowth syndrome
Porencephaly	Platelets; Coagulation profile (PT, INR, PTT); Consider additional haematological work-up; Hereditary Cerebral Small Vessel Diseases Panel, exome or genome sequencing; COL4A1, COL4A2 related work-up including ophthalmological exam (by neuro-ophthalmology where available), serum creatine kinase, ECG, renal ultrasound
Schizencephaly	Monitor glucose and electrolytes; rule out hypopituitarism; ophthalmological exam (by neuro-ophthalmology where available) for optic nerve hypoplasia; Hereditary Cerebral Small Vessel Diseases Panel, exome or genome sequencing; consider COL4A1, COL4A2 related work-up
Polymicrogyria	Consider 22q11.2 deletion syndrome work-up if bilateral perisylvian or associated with ACC or cerebellar anomalies; consider infectious and metabolic studies if other features as below; Brain Malformation Panel, exome or genome sequencing; consider COL4A1, COL4A2 related work-up if focal
Selected patterns
Microcephaly, ventriculomegaly, polymicrogyria, calcification and/or cerebellar anomalies (suggestive of congenital infection)	Examine for systemic findings of rash or hepatosplenomegaly; if CMV negative, serology and/or PCR for other infections (toxoplasmosis, parvovirus, varicella, syphilis, rubella, herpes, Zika virus, lymphocytic choriomeningitis virus); CBC; liver enzymes; abdominal ultrasound; ophthalmological examination (by neuro-ophthalmology where available); long bone radiographs; audiology
Polymicrogyria, pachygyria, or simplified gyri with T2-weighted hyperintensity of the white matter +/− periventricular/germinolytic cysts (suggestive of peroxisomal disorders)	Review newborn screen results; very long chain fatty acids; liver enzymes, renal function; abdominal ultrasound; skeletal survey for chondrodysplasia punctata; ophthalmological examination (by neuro-ophthalmology where available); audiology
Cobblestone malformation and hindbrain malformation (suggestive of α-dystroglycanopathy)	Creatine kinase; ophthalmologic exam (by neuro-ophthalmology where available); ECG and echocardiogram; neuromuscular panel, exome or genome sequencing

ACC = anterior cingulate cortex; CBC = complete blood count; CMV = congenital cytomegalovirus; PT = prothrombin time; INR = international normalized ratio; PTT = partial thromboplastin time.

Neurodevelopmental follow-up and support

Serial examinations should address developmental progress while continuing to reassess and reevaluate any diagnostic uncertainty; for example, repeated dermatologic exams for emergence of pathognomonic neurocutaneous findings¹¹¹ or periodic reanalysis of uncertain or non-diagnostic genetic testing results. A suggested schedule mirroring neurodevelopmental assessment in other high-risk populations with follow-up at 3 months, 6 months, 12 months, 24 months, preschool age and then at school age is recommended, with more frequent visits as needed when there are neurodevelopmental impairments or other neurological diagnoses requiring management, such as epilepsy and cerebral palsy.

All infants with MCDs should be enrolled in outpatient early intervention therapies that include physical, occupational, feeding, speech and/or vision therapy as appropriate to support developmental progress and address equipment needs to ensure safety and inclusion.

Research gaps and future directions

The advancement of genetic testing and fetal imaging is improving our ability to detect MCDs, uncover their aetiologies and provide individualized patient care. However, to continue to improve detection, prognostication and outcomes for children with MCDs, future research is needed to address numerous current knowledge gaps. First, additional imaging studies will be required to optimize fetal detection of MCDs and evaluate the predictive value of fetal MRI for unique MCD subtypes. For instance, the predictive value of non-specific signs on fetal imaging, such as abnormal enlargement or cavitation of the GEs beyond 34 gestational weeks, is not known, although such signs can precede diagnoses of MCDs. Studies that systematically correlate fetal neuroimaging with postnatal imaging, or with the gold standard of post-mortem tissue evaluation when available, in larger multicentre cohorts would be ideal for honing the predictive value of fetal imaging for distinct MCD subtypes. Second, longitudinal neurosonogram and fetal MRI studies can also improve our understanding of how secondary MCDs evolve over time, which may help refine subtle features of both diagnosis and prognosis for distinct subtypes of MCDs and help distinguish them from primary MCDs. Finally, novel research approaches to fetal brain imaging, including diffusion tensor imaging, slice to volume registration, MR spectroscopy and functional MRI, can provide further insights into the underlying mechanisms of prenatal cortical development.^112-115

Since the spectrum of long-term outcomes for many MCDs can be wide and difficult to predict, future studies will also be necessary to help refine prognoses and stratify fetuses by their risk for developmental domain-specific delays and impairments, epilepsy and cognitive and/or learning disabilities. Understanding the short- and long-term risk for developing infantile spasms and other forms of epilepsy across developmental stages by MCD type may also impact our approach to obtaining screening EEGs for select high-risk constellations of MCDs prior to seizure onset. In studies of patients with TSC,^116-118 it is hypothesized that earlier detection of epileptiform activity and preventive initiation of anti-seizure medications can reduce long-term seizure burden and improve cognitive outcomes; future research should investigate whether these findings can be generalized to other MCDs. Partnering with children and their families to help monitor for and report neurodevelopmental symptoms will be essential to successfully carrying out longitudinal outcome studies.

Basic and translational scientific advancements will help improve care for patients with MCDs. Developing novel methods of genetic screening and testing and advanced quantitative imaging may improve timely detection of MCDs during fetal development. Understanding the cell-specific pathophysiology of MCDs, particularly the selective vulnerability of specific cortical cell types to secondary injury at distinct gestational ages and then correlating this data with the anatomy and phenotype of MCD subclasses, will improve our prognostic abilities. Such work may even suggest innovative strategies for preventing formation of acquired MCDs after in utero insult. Last, the development of innovative genetic and molecular therapeutics to treat MCDs, which will likely rely on cutting edge in utero treatment strategies,^119,120 will help improve outcomes for patients with MCDs.

Conclusions

Malformations of cortical development are a broad family of disorders that result in abnormal development of the fetal cerebral cortex. Despite their heterogeneity, as a class, they represent some of the most commonly detected prenatal congenital brain malformations. MCDs lead to a high burden of paediatric morbidity, including epilepsy and neurodevelopmental impairment, including intellectual disability, sensory impairment and cerebral palsy, as well as a higher risk for early mortality. While our ability to detect MCDs antenatally has improved with more nuanced fetal imaging and antenatal genetic testing, there are few formal consensus guidelines around (i) optimal timing, frequency and modality of fetal imaging to detect and monitor MCDs; (ii) delivery and early postnatal management; and (iii) best practices to reduce comorbidities and optimize developmental outcomes for children with MCDs. In this review, we summarize the existing literature on perinatal detection and management of MCDs and provide consensus practical recommendations for paediatric neurologists. Standardization of care will facilitate long-term natural history studies and registry development. Future research is essential to optimize antenatal imaging algorithms, refine prognostication, address disparities in care and develop innovative therapeutics for patients with MCDs.

Funding

This work was supported in part by National Institutes of Health grants 1K08NS133292 (J.B.R.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Competing interests

The authors report no competing interests.

Supplementary material

Supplementary material is available at Brain online.