Spectrum of Fetal Brain Anomalies Depicted on Fetal MRI

Fetal MRI is now offered as the standard of care in appropriate clinical settings, proving its worth in exquisite soft tissue detail. Sonographic fetal anatomic survey remains the screening mainstay for cost, comfort, and availability. However, fetal MRI has added utility in questionable cases, the search for additional findings with known abnormalities, and screening at-risk fetuses.¹ Fetal MRI has emerged as a complement to sonographic fetal anatomic survey without adverse effects observed.² Benefits of fetal MRI on 1.5 tesla (T) magnet outweigh the risks in clinically indicated studies.³

Routinely, fetal MRI is performed on a 1.5T magnet in the second or third trimester. First trimester fetuses are not imaged with MRI. No gadolinium contrast agents are administered as they cross the placenta.² Some institutions perform fetal MRI on a 3T system, which provides superior signal to noise ratio, but requires management of numerous artifacts. Fetal 3T MRI is not yet routine and may require consent.³

During imaging, the mother is placed in the position of most comfort (usually left lateral decubitus or supine). Both maternal sedation and NPO requirements are highly variable; many institutions do not use either. A multichannel, phased array coil is placed low on the maternal abdomen to maximize coverage of the fetus. Initial localizer images of the mother are obtained to determine fetal position. Planes based on orientation of the fetus are then planned. T2-weighted imaging using ultrafast sequences (single shot fast spin echo) are mainly used.¹ T1-weighted images (both with and without fat saturation), diffusion weighted imaging, and gradient images are often obtained for complementary information.

Referrals for MRI evaluation of the fetal brain are most commonly secondary to enlarged ventricles, callosal dysgenesis, and posterior fossa abnormalities, as well as complications of monochorionic twinning (not discussed here).¹ There are limitations of fetal MRI, however. For example, the full spectrum of cortical malformation may not be evident due to the limited spatial resolution and relative lack of myelination. White matter diseases are not detectable. Postnatal MRI follow-up may be appropriate as ongoing myelination and improved spatial resolution can reveal more findings.

Supratentorial Brain

The MRI appearance of the fetal brain must be correlated with an accurate gestational age, usually achieved by first trimester dating ultrasound and clinical estimation from the last menstrual period. Knowledge of the unique multilayered pattern of the fetal brain and progressive sulcal and gyral development is essential for assessment.

The supratentorial brain has a 3-layer pattern on MRI in early second trimester gestation. These layers are ventricular zone/germinal matrix (dark T2 signal), intermediate zone (high T2 signal), and cortical plate (low T2 signal).⁴ A 5-layered pattern is reliably seen between 20-28 weeks gestation.⁵ The 5 layers from medial to lateral are the ventricular zone/germinal matrix (dark T2 signal), periventricular zone (high T2 signal), intermediate zone (low T2 signal), subplate (high T2 signal), and cortical zone (dark T2 signal). Zones are best depicted on T2-weighted images where the cortex is a dark rim (Figure 1). The zones reflect migrating cells toward the cortex from the germinal matrix.⁴ Eventually, the multilayered appearance matures to the typical postnatal brain MR appearance.⁵

There is progressive, orderly appearance of the sulci associated with predictable gestational ages in normal brain development. When reviewing the literature it is important to ascertain whether the description of visualized sulci is from pathology, postmortem MRI or fetal MRI, as pathologic specimens demonstrate sulci 2-4 weeks earlier than fetal MRI. Sylvian fissures are identified on fetal MRI before 18 weeks of gestation.⁵ Between 22-25 weeks, more than 75% of normal fetal brains have callosal, parietoccipital, cingulate, calcarine, and hippocampal fissures on MRI. At the vertex, the central, precentral, and postcentral sulci are seen between 27-28 weeks on MRI. Superior frontal, inferior frontal, superior temporal, and inferior temporal sulci are present by 29-34 weeks.⁶ The corpus callosum forms between 8-20 weeks and can be seen on midline sagittal MR image by 20 weeks.^5,7

The authors recommend the text by Griffiths PD et al as a general reference for normal fetal and postnatal brain MRI appearance at all stages of development.⁸ Lack of knowledge regarding prenatal brain maturation (multilayered appearance and expected sulcation) are considered a pitfall of fetal MRI. As discussed above, correlation with accurate gestational age is also essential.⁹

Supratentorial Pathology

Ventriculomegaly is a common indication for second trimester fetal MRI. Ventriculomegaly is graded as mild (> 10-12 mm), moderate (> 12-15 mm) and severe (> 15 mm).¹⁰ Ventricles are measured at the level of the atrium, typically on axial or coronal images. The MR and sonographic measurements can vary by 1-2 mm on the axial images. Coronal image measurements are more concordant at the level of atria and mid-height of the ventricle.^1,11 Subtle asymmetry of ventricular size without overt ventriculomegaly is a normal variant. Mild to moderate ventriculomegaly in isolation is associated with a better, often normal, neurodevelopmental outcome. However, up to 50% of fetuses with ventriculomegaly on ultrasound show additional CNS findings on MRI, such as agenesis corpus callosum, hemorrhage, encephalomalacia, heterotopia, and posterior fossa abnormalities.¹ Severe ventriculomegaly must be differentiated from hydranencephaly and holoprosencephaly.

Callosal anomalies can range from agenesis to hypoplasia. Absence of the cavum septum pellucidum is seen with callosal agenesis, but also holoprosencephaly. Parallel configuration of the lateral ventricles with a “steer horn” configuration of the frontal horns is characteristic of agenesis of the corpus callosum (Figure 2), as is colpocephaly, which refers to disproportionate enlargement of the occipital horns of the lateral ventricles (Figure 3). Midline lipoma and interhemispheric cysts are associated with corpus callosal abnormalities (Figure 4). Additional brain anomalies are seen on MRI in more than 75% of fetuses with callosal agenesis, including abnormal cortical mantle (Figure 4), posterior fossa abnormalities including vermian dysgenesis, and brain stem hypoplasia.^1,7 Severe ventriculomegaly limits evaluation of the corpus callosum, as it can be extremely thinned.

Holoprosencephaly is a spectrum of disorders characterized by incomplete cleavage of the prosencephalon, which forms the telencephalon (forms the cortex) and the diencephalon (forms the thalami). The spectrum ranges from absence of these structures in aprosencephaly to alobar, semilobar, and lobar holoprosencephaly variants. Classic alobar holoprosencephaly is characterized by absence of midline cleavage leading to contiguous frontal lobes, fused thalami and basal ganglia, and an underlying mono ventricle. Semilobar is less severe with some cleavage, usually posteriorly, and variable separation of thalami (Figure 5). Septo-optic dysplasia is often considered in the spectrum of holoprosencephaly as the least severe form.¹²

Lissencephaly can be diagnosed on fetal MRI after 20 to 22 weeks gestation. ^{13, 14} Correlation with accurate gestational age and absence of expected sulci is crucial, as the normal fetal brain has a smooth contour earlier in gestation (“physiologic lissencephaly”).¹⁴ Lissencephaly is due to impaired migration of neurons early in gestation (third or fourth month). It manifests as smooth brain with absence (agyria) or abnormal appearance (pachygyria) of the expected sulci and gyri (Figure 6). It is associated with many syndromes, including Miller-Dieker in classic type and Walker-Warburg in cobblestone type. Delayed and abnormal sulcation is associated with a wide variety of CNS anomalies other than lissencephaly, however, including ventriculomegaly (which can negate evaluation of the sulci in more severe cases with cortical thinning), holoprosencephaly, and extensive hemorrhage.¹³

Periventricular nodular heterotopias are visible on prenatal MRI as hypointense regions on T2-weighted sequences protruding into the ventricular lumen. Sensitivity of prenatal MRI for detecting periventricular nodular heterotopia is reported to be between 40% (prospectively) to 73% (retrospectively) and with similar specificities of 91% to 92%. When seen in 2 planes, specificity increases to 100%. There is lower sensitivity for detection of periventricular nodular heterotopia in fetuses at < 24 weeks of gestation mainly due to relative thicker germinal matrix, which has the same signal intensity as periventricular heterotopia. No susceptibility is seen on gradient images in heterotopia, which helps distinguish it from hemorrhage. Almost all cases of cortical malformation in one review (including periventricular nodular heterotopia) had additional abnormalities detected on fetal MRI.¹⁵

Polymicrogyria manifests as focal (Figure 7) or diffuse nodularity along the cortical surface arising from abnormal infoldings of the developing cortex. Polymicrogyria can be associated with underlying abnormal subplate and intermediate zone.¹⁴ One study reported fetal MRI to have 85% sensitivity for detecting polymicrogyria.¹⁵ Etiologies of polymicrogyria include intrauterine infection, vascular insult, and genetic etiologies, as well as syndromic associations, some of which are familial.¹⁶

Schizencephaly manifests as continuation of abnormally sulcated gray matter signal along a cerebrospinal fluid (CSF) cleft from the ventricular surface to the pial surface. Most are associated with an absent septum pellucidum. Schizencephaly is classified into closed or open lip types, depending on whether the cleft surfaces are opposed to each other (closed) or separated by a pocket of CSF (open). The open lip type may be covered by a membrane, particularly in fetal life. The adjacent cortex is abnormal and typically demonstrates polymicrogyria. There is usually adjacent focal dilation of the ventricle.¹⁵ One study demonstrated that nearly half of the cases of open lip schizencephaly can close by birth. The location of the cleft and unilateral vs. bilateral involvement have prognostic significance.¹⁷

The most common intracranial cysts are choroid plexus cysts and arachnoid cysts. When seen in isolation, both have a favorable outcome. Multiple choroid plexus cysts and other sonographic markers of aneuploidy can be seen in Trisomy 18. Arachnoid cysts are extra-axial and can be associated with agenesis of the corpus callosum when interhemispheric in location (Figure 4). Intraparenchymal cystic space lined by gliotic tissue (porencephalic cyst) results from a prenatal infectious or vascular insult and can communicate with the lateral ventricle. Unlike schizencephaly, porencephalic cysts are not lined by dysplastic gray matter. Periventricular pseudocysts (germinolytic cysts) occur at the caudothalamic groove in a variety of infectious, noninfectious, metabolic, and chromosomal abnormalities. Caudothalamic groove cysts can also be due to germinal matrix hemorrhage.¹⁸

Intracranial hemorrhage can be intraventricular, periventricular, intraparenchymal or extra-axial. Predisposing factors range from maternal bleeding disorders and infections, to fetal chromosomal and metabolic abnormalities. Hemorrhage can have varying signal characteristics, but often is dark on T2 images with associated susceptibility. Germinal matrix hemorrhage is often asymmetric and irregular with associated susceptibility compared to the normal dark T2 signal of the germinal matrix. Hemorrhage along the ventricles can evolve into periventricular cystic leukomalacia (Figure 8). Larger hemorrhages are associated with poor neurodevelopmental outcomes.

Among vascular abnormalities, vein of Galen malformation and dural venous sinus thrombosis deserve mention. Vein of Galen malformation is an arteriovenous fistula, and a common cause of high-output heart failure. The fistulous arteries connect to the persistent median prosencephalic vein of Markowski, which should regress before the 11th week of gestation. This persistent connection results in dilation of the vein of Galen. Cerebral vascular steal leads to brain ischemia and hydrocephalus. Sonography with color Doppler is sufficient for diagnosis, although fetal MRI can add detail regarding brain parenchyma.¹⁹ Dural venous sinus thrombosis may be idiopathic or associated with maternal preeclampsia, hypercoagulation, or infection. On MRI, a large extra-axial mass due to thrombus is typically iso to hypointense to gray matter with a more focal T2 hypointense focus (Figure 9). Rarely, dural sinus thromboses arise from an underlying dural arteriovenous fistula and have a worse prognosis.²⁰

Bilateral internal carotid artery territory destruction leading to hydranencephaly is one of the most severe insults to the brain. Most of the supratentorial parenchyma is replaced by fluid covered meninges. No peripheral rim of cortex is present. The brainstem, cerebellum, portions of basal ganglia, thalami, and medial occipital cortex are spared as the posterior circulation territory is preserved (Figure 10). Fetal MRI allows for a more precise diagnosis by differentiating hydranencephaly from severe hydrocephalus and alobar holoprosencephaly (both of which have a thin preserved cortical mantle even in the most severe forms). Hydranencephaly is associated with a short postnatal survival in most cases. ²¹

Multiple in utero infections can have brain manifestations, especially TORCH (Toxoplasmosis, Other [syphilis, Varicella-Zoster], Rubella, Cytomegalovirus, Herpes) infections. Fetal cytomegalovirus infection can lead to periventricular calcification, polymicrogyria, microcephaly, porencephaly, white matter signal abnormality, and cerebellar hypoplasia. Diffuse calcifications, cortical dysplasia, and parenchymal destructive lesions are associated with fetal toxoplasmosis, rubella, and Herpes simplex infections. However, calcifications are difficult to identify prenatally. Lissencephaly and hydrocephalus can be seen with fetal varicella infection. Fetal infections are difficult to distinguish radiologically from a pseudo TORCH syndrome/Aicardi-Goutiere syndrome, which is a familial autosomal recessive disease with similar cortical malformations and calcification.²²

Posterior Fossa

Indications for fetal MRI of the posterior fossa most often relate to abnormality in cerebellar size (including vermis) and enlargement of the cisterna magna. Posterior fossa evaluation should include the size of the cisterna magna, insertion of the tentorium, size and morphology of the cerebellum and vermis, tegmento-vermian angle (TVA), shape of the fourth ventricle, morphology of the pons, and the supratentorial brain for any associated abnormalities.²

The normal size of the cisterna magna should be < 10 mm on axial images.^2,23 The tentorium should insert at the level of the torcula and course perpendicular to the occiput.²

The transverse cerebellar diameter (TCD) is a reliable measurement to assess cerebellar size based on gestational age. It is unaffected by other fetal abnormalities, such as growth restriction.²⁴ TCD is measured on an axial image across the widest dimension of both cerebellar hemispheres. TCD in mm closely approximates the gestational age in weeks from 19-25 weeks.^25,26

The cerebellar vermis develops in the roof of the precursor to the fourth ventricle (rhombencephalic vesicle).²³ The vermis forms from proliferation of midline tissue. A nearly equal amount of vermian tissue is seen above (47%) and below (53%) the midpoint (fatigium declive line) with linear growth during gestation. The fastigial point and primary fissure of the vermis should be visualized by 18 weeks of gestation on MRI (Figure 11); however there can be physiologic delay up to 24 weeks.^23,27 Vermian measurements in the anteroposterior and craniocaudal dimensions on a midsagittal image should be correlated to normative data.²⁵ Foliation of the vermis increases with increasing gestational age,^{2, 25-27} and better lobulation correlates to improved prognosis. At the earliest, 9 lobules can be identified at 27 weeks of gestation.²⁷

The TVA is calculated between a line along the posterior surface of the medulla parallel to the tegmentum and a line along the anterior surface of the vermis. Normally, both lines are nearly vertical with an angle close to 0 degrees. Angles between 0-40 degrees can be due to a normal but rotated vermis (ie, Blake’s pouch cyst) and vermian hypoplasia (Figure 12). When > 40 degrees, there is high association with vermian hypoplasia.²⁷ The normal fourth ventricle has a triangular shape in the sagittal plane and is covered by the vermis by 18 weeks, but no later than 24 weeks of gestation.²³

Vermian hypoplasia is often overdiagnosed on fetal MRI and has an uncertain prognosis when present in isolation. In one study, prenatal diagnosis of vermian hypoplasia on MRI was confirmed postnatally in only 68% of patients. In addition, 77% patients with confirmed vermian hypoplasia postnatally had a normal neurologic outcome.²⁸ Follow-up MRI should be pursued to confirm the diagnosis.² Evaluation of the vermis in all 3 planes is urged.¹ Evaluating the sagittal plane alone can, for example, overestimate the amount of vermian tissue present from overlap of the cerebellum. Imaging findings include a small vermis, abnormal foliation, increased TVA, normal-sized posterior fossa, and normal insertion of the tentorium (Figure 12).^25,27

Along the undersurface of the vermis at the inferior roof of rhombencephalic vesicle, there is a developmental outpouching. This diverticulum, Blake’s pouch, is lined with ependyma and has choroid plexus in its roof. It communicates with the fourth ventricle and contains intra-axial CSF. Blake’s pouch is a normal, transient structure that fenestrates to communicate with the subarachnoid space of cisterna magna. This fenestration and opening of the foramen of Luschka help distribute CSF normally throughout the posterior fossa. Cisterna magna septa, coursing straight from the lateral margin of the vermis to occiput, are likely walls of Blake’s pouch and normal structures, although variably seen (Figure 13).²³

Dandy Walker continuum is a spectrum of abnormalities related to the roof of rhombencephalic vesicle, including mega cisterna magna, Blake’s pouch cyst, and classic Dandy Walker malformation.²³ Outcomes are highly variable from normal cognition to severely affected (syndromes) and are correlated to the degree of vermian hypoplasia and other abnormalities present (ie, pontine hypoplasia). There are genetic associations such as Walker Warburg and Meckel-Gruber syndromes.²⁷

Mega cisterna magna is defined as increased CSF space (> 10 mm) posterior to a normally formed cerebellum (Figure 13).² The proposed etiology is dilation of Blake’s pouch due to delayed, but eventual fenestration, resulting in an expanded posterior fossa subarachnoid space.²³

Blake’s pouch cyst communicates with the fourth ventricle and exerts mass effect on the vermis. The vermis is normally formed; however, it is rotated and can appear hypoplastic. Blake’s pouch cyst is likely due failure of fenestration and inadequate opening of the foramina of Luschka. Therefore CSF within Blake’s pouch cannot equilibrate into the subarachnoid space of the posterior fossa.²³ Imaging findings include a CSF filled cyst beneath the vermis that communicates with the fourth ventricle with choroid plexus extending from the fourth ventricle into the cyst. This can be associated with hydrocephalus.²⁹

Classic Dandy Walker malformation includes vermian agenesis or hypoplasia, superiorly displaced tentorium, enlarged posterior fossa, and communication of an enlarged posterior fossa CSF collection with the fourth ventricle.² Imaging findings include small or absent vermis (lack of primary fissure and fastigial point), increased TVA, high tentorium, and enlarged posterior fossa (Figure 14).

Vermian agenesis has a differential that includes Dandy Walker malformation, rhombencephalosynapsis, and Joubert syndrome. Rhombencephalosynapsis is characterized by fusion of cerebellar hemispheres across the midline with vermian agenesis. It is due to a failure of midline division and associated with holoprosencephaly.² Joubert syndrome is due to lack of decussation of superior cerebellar peduncles, corticospinal tracts, and central pontine tracts. Pathognomonic “molar tooth” finding on axial images at the midbrain, thickened superior cerebellar peduncles, absent vermis, and a deformed fourth ventricle are the characteristic findings (Figure 15).³⁰ Like vermian hypoplasia, agenesis must be assessed in all planes. In the axial plane, abnormal opposition of the cerebellar hemispheres yields a cleft. TCD can be normal or increased in vermian agenesis.²

A posterior fossa arachnoid cyst is in the differential for many of these entities. However, it is extra-axial and will not communicate with the fourth ventricle. It exerts mass effect on the posterior fossa structures and often is off of midline.²

A small posterior fossa can be seen in Chiari II resulting from a lumbosacral myelomeningocele (Figure 16). The cerebellum is herniated inferiorly through the foramen magnum with obliteration of the CSF space surrounding the cerebellum and a low tentorium.² This often results in supratentorial hydrocephalus. Fetal MR imaging is used to assess the degree of hindbrain herniation, ventriculomegaly, and the level of spinal dysraphism. Associated supratentorial abnormalities seen in patients with myelomeningocele include corpus callosal hypoplasia (57% patients), heterotopia, and polymicrogyria.¹ The fetal spine can be assessed on the MR images obtained though the fetal abdomen. The ending of the conus and open spinal dysraphism are identifiable. A fatty filum, tethered cord, and closed spinal dysraphism are not reliably assessed on fetal MRI.

Conclusion

Fetal MRI is a valuable adjunct for clinicians and patients, often confirming sonographic findings, revealing additional abnormalities, and ultimately aiding in appropriate patient counseling. The radiologist must be well-versed in normal brain development and have an accurate gestational age in order to correctly interpret fetal MRI. The spectrum of detectable anomalies (many of which are reviewed here) is vast, but not all-inclusive. Therefore, consideration to postnatal MRI follow-up in appropriate settings (ie, vermian hypoplasia) remains extremely important.

References

1. Glenn O. MR Imaging of the fetal brain. Pediatr Radiol 2010;40:68-81.
2. Adamsbaum C, Moutard M, Andre C, et al. MRI of the fetal posterior fossa. Pediatr Radiol 2005;35:124-140.
3. Victoria T, Jaramillo D, Roberts T, et al. Fetal magnetic resonance imaging: jumping from 1.5 to 3 Tesla (preliminary experience). Pediatr Radiol 2014;44:376-386.
4. Griffiths PD, Morris J, Larroche JC, et al. Sectional anatomy of the fetal brain. In: Griffiths PD, Morris J, Larroche JC, Reeves M, eds. Atlas of Fetal and Postnatal Brain MRI. Philadelphia, PA: Mosby, Elsevier; 2010.
5. Glenn OA, Barkovich AJ. Magnetic resonance imaging of the fetal brain and spine: an increasingly important tool in prenatal diagnosis, part 1. Am J Neuroradiol 2006;27:1604-1611.
6. Garel C, Chantrel E, Brisse H, et al. Fetal cerebral cortex: normal gestational landmarks identified using prenatal MR imaging. Am J Neuroradiol 2001 22(1):184-189.
7. Glenn O, Goldstein RB, Li KC, et al. Fetal MRI in the evaluation of fetuses referred for sonographically suspected abnormalities of the corpus callosum. J Ultrasound Med 2005;24:791-804.
8. Griffiths PD, Morris J, Larroche JC, et al. Atlas of Fetal and Postnatal Brain MRI. Philadelphia, PA: Mosby, Elsevier; 2010.
9. Al-Mukhtar A, Kasprian G, Schmook MT, et al. Diagnostic pitfalls in fetal brain MRI. Semin Perinatol 2009;33(4):251-258.
10. Pilu, G. Ultrasound evaluation of the fetal neural axis. In: Callen PW, ed. Ultrasonography in Obstetrics and Gynecology. 5th ed. Philadelphia, PA: Saunders, Elsevier; 2008.
11. Garel C, Alberti C. Coronal measurement of the fetal lateral ventricles: comparison between ultrasonography and magnetic resonance imaging. Ultrasound Obstet Gynecol 2006;27(1):23-27.
12. Winter TC, Kennedy AM, Woodward PJ. Holoprosencephaly: a survey of the entity, with embryology and fetal Imaging. Radiology 2015;35(1):275-290.
13. Ghai S, Fong K, Toi A, et al. Prenatal US and MR imaging findings of lissencephaly: review of fetal cerebral sulcal development. Radiographics 2006;26(2):389-405.
14. Righini A, Parazzini C, Doneda C, et al. Early formative stage of human focal cortical gyration anomalies: fetal MRI. Am J Roentgenol 2012;198(2):439-447.
15. Glenn OA, Cuneo AA, Barkovich AJ, et al. Malformations of cortical development: diagnostic accuracy of fetal MR imaging. Radiology 2012;263(3):843-855.
16. Squier W, Jansen A. Polymicrogyria: pathology, fetal origins and mechanisms. Acta Neuropathol Commun 2014; 2:80.doi:10.1186/s40478-014-0080-3.
17. Nabavizadeh SA, Zarnow D, Bilaniuk LT, et al. Correlation of prenatal and postnatal MRI findings in schizencephaly. Am J Neuroradiol 2014;35(7):1418-1424.
18. Pappalardo EM, Militello M, Rapisarda G, et al. Fetal intracranial cysts: prenatal diagnosis and outcome. J Prenatal Med 2009;3(2):28-30.
19. Messori A, Polonara G, Salvolini, U, et al. Prenatal diagnosis of a vein of Galen aneurysmal malformation with fetal MR imaging study (multiple letters). Am J Neuroradiol 2003;24(9),1923-1925.
20. Byrd SE, Abramowicz JS, Kent P, et al. Fetal MR imaging of posterior intracranial dural sinus thrombosis: a report of three cases with variable outcomes. Pediatr Radiol 2012;42(5):536-543.
21. Ghosh PS, Reid JR, Patno D, et al. Fetal magnetic resonance imaging in hydranencephaly. J Paediatr Child Health 2013;49(4):335-336.
22. Sreedher G, Panigrahy A. Central nervous system infections. In: Bahado-Singh R, ed. Fundamental and Advanced Fetal Imaging. Ultrasound and MRI. Philadelphia, PA: Wolters Kluwer; 2015.
23. Robinson A, Goldstein R. The cisterna magna septa: vestigial remnants of Blake’s pouch and a potential new marker for normal development of the rhombencephalon. J Ultrasound Med 2007;26(1):83-95.
24. Davies M, Swaminathan M, Betheras F. Measurement of the transverse cerebellar diameter in preterm neonates and its use in assessment of gestational age. Australas Radiol 2001;45:309-312.
25. Triulzi F, Parazzini C, Righini A. MRI of fetal and neonatal cerebellar development. Semin Fetal Neonatal Med 2005;10:411-420.
26. Triulzi F, Parazzini C, Righini A. Magnetic resonance imaging of fetal cerebellar development. Cerebellum 2006;5:199-205.
27. Robinson A, Blaser, S, Toi, A, et al. The fetal cerebellar vermis: assessment for abnormal development by ultrasonography and magnetic resonance imaging. Ultrasound Q 2007;23(3):211-223.
28. Limperopoulos C, Robertson R, Estroff J, et al. Diagnosis of inferior vermian hypoplasia by fetal MRI: potential pitfalls and neurodevelopmental outcome. Am J Obstet Gynecol 2006;194(4):1070-1076.
29. Cornips EM, Overvliet GM, Weber JW, et al. The clinical spectrum of Blake’s pouch cyst: report of six illustrative cases. Child’s Nerv Syst 2010;26:1057-1064.
30. McGraw P. The molar tooth sign. Radiology 2003;229(3):671-672.