PRENATAL DIAGNOSIS

Prenat Diagn 2009; 29: 372–380. Published online 4 February 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/pd.2196

REVIEW

The fetal cerebellum. Pitfalls in diagnosis and management Gustavo Malinger*, Dorit Lev and Tally Lerman-Sagie Fetal Neurology Clinic, Prenatal Diagnosis Unit, Department of Obstetrics and Gynecology; Genetics Institute and Pediatric Neurology Unit, The Edith Wolfson Medical Center, Holon and Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel Prenatal diagnosis of congenital and acquired cerebellar disorders is possible by the use of ultrasound (US) and magnetic resonance imaging (MRI). Although numerous studies have been conducted in this field, diagnostic uncertainties are still common in daily clinical practice. This review outlines five possible pitfalls in the diagnosis of fetal cerebellar disorders: confusion between different entities describing vermian pathologies (Dandy–Walker variant, vermian hypoplasia and vermian agenesis); premature diagnosis of abnormal vermian formation; difficulties in the ultrasonographic differentiation between the cerebellar hemispheres and the vermis; late development of cerebellar hypoplasia/atrophy and differential diagnosis of unilateral cerebellar findings. Copyright  2009 John Wiley & Sons, Ltd. KEY WORDS:

fetal cerebellum; cerebellar congenital malformations; fetal ultrasound

More than 100 years ago, in 1898, Joseph A. Blake read a paper before the Association of American Anatomists on the development of the fourth ventricle. His opening words were: My investigations on this subject were prompted by the contradictory opinions and lack of absolute knowledge concerning the nature of the communications between the cavity of the fourth ventricle and the subarachnoid space (Blake, 1900) Despite his and other investigators studies and advances in imaging, including the introduction of ultrasound and magnetic resonance; our knowledge on cerebellar development beyond the embryologic period remains limited; precluding in many cases clear differentiation between normal and pathological conditions. Carroll et al. (2000) found a lack of correlation in 57% (8 out of 14 fetuses) when comparing the results of the fetal ultrasound examinations in which, a cerebellar anomaly was diagnosed to the results of autopsies on the aborted fetuses. Similar results were recently obtained in a larger study of 44 fetuses with an ultrasound (US) diagnosis of Dandy–Walker malformation (DWM); the US examination failed to correctly diagnose DWM in 26 patients (Phillips et al., 2006). The introduction of magnetic resonance as a complementary method of diagnosis has shown that a correct diagnosis of inferior vermian hypoplasia is difficult even with this technique with a false positive rate of 32% (Limperopoulos et al., 2006). Inferior vermian hypoplasia has been associated with many mental retardation syndromes and with autism; therefore, fetal medicine experts should be aware and extremely careful when making this diagnosis in utero. *Correspondence to: Gustavo Malinger, Prenatal Diagnosis Unit, Department of Obstetrics and Gynecology, The Edith Wolfson Medical Center, Holon 58100, Israel. E-mail: [email protected]

Copyright  2009 John Wiley & Sons, Ltd.

Potential pitfalls also exist in the diagnosis of fetuses with cerebellar hypoplasia. The purpose of this review is to present and discuss the different ultrasonographic patterns that may lead to a false positive or negative diagnosis in the context of normal and abnormal fetal cerebellar development. An approach to the evaluation of the normal cerebellum and vermis and different flow charts for the diagnosis of specific pathologies have been previously suggested by different groups (Smith et al., 1986; Pilu et al., 1987; Malinger et al., 2001; Adamsbaum et al., 2005; Guibaud and des Portes, 2006; ISUOG, 2007). The ultrasonographic evaluation should include multiplanar images of the cerebellum (Figure 1). The axial plane is useful for determining the transcerebellar diameter, fourth ventricle size and the cerebellar peduncles thickness (Figure 1A). The coronal plane enables differentiation between the cerebellar hemispheres and the vermis (Figure 1B). The midsagittal plane is the most important plane to be investigated since it allows depiction of the vermian lobules, fissures and shape of the fastigium; measurements of the vermian diameter, surface and calculation of the ratio between the superior and inferior parts. This plane also permits evaluation of the size and shape of the pons, cisterna magna and tentorium (Figures 1C and 2E). On the basis of the experience accumulated in the Fetal Neurology Clinic at Wolfson Medical Center on 125 fetuses with suspected cerebellar anomalies (Halawa et al., 2007), we outline five different situations that may result in diagnostic and counseling errors: confusion between different entities describing vermian pathologies (Dandy–Walker variant, vermian hypoplasia and vermian agenesis); premature diagnosis of abnormal vermian formation; difficulties in the ultrasonographic differentiation between the cerebellar hemispheres and the vermis; late development of cerebellar hypoplasia/atrophy and differential diagnosis of unilateral cerebellar findings. Received: 26 July 2008 Revised: 9 October 2008 Accepted: 21 November 2008 Published online: 4 February 2009

THE FETAL CEREBELLUM

373

Figure 1—Multiplanar evaluation of the cerebellum. (A) Axial plane at 28 weeks with the calipers measuring the transcerebellar diameter (cereb) and the width of the posterior fossa (2); (B) coronal plane at 33 weeks, the vermian fissures are shown as echogenic parallel lines (arrows). The cerebellar hemispheres are less echogenic; superficial planes may enable visualization of the cerebellar folia (arrowed); (C) midsagittal plane at 33 weeks showing the vermian lobules, fissures and shape of the fourth ventricle with the fastigium (F). Other structures depicted are the cisterna magna (CM), pons (P), tentorium (T) and corpus callosum (CC)

The agreed definitions of the different cerebellar pathologies that may be diagnosed in utero are presented in Table 1. CONFUSION BETWEEN DIFFERENT ENTITIES DESCRIBING VERMIAN PATHOLOGIES (DANDY–WALKER VARIANT, VERMIAN HYPOPLASIA AND VERMIAN AGENESIS) The confusion in the literature on vermian abnormalities stems from two main problems: the use of the term Dandy–Walker variant (DWV) that includes a heterogeneous group of disorders and inaccurate use of the terms vermian agenesis and hypoplasia. Parisi and Dobyns (2003) in their article on cerebellar and hindbrain malformations deal with the confusion created by the use of the term DWV for the description of different degrees of cerebellar vermis hypoplasia with slight or absent upward rotation of the vermis, and various amounts of posterior fossa fluid collections. They recommend abandoning this term given its variable definitions, lack of specificity and confusion with classic DWM. When reviewing the literature, particularly studies dealing with the prenatal diagnosis of cerebellar malformations, the term DWV is used synonymously with Copyright  2009 John Wiley & Sons, Ltd.

isolated vermian hypoplasia, inferior vermian hypoplasia and even with vermian dysgenesis/agenesis as part of molar tooth related syndromes (MTRS). The use of this inaccurate terminology led to the finding of different outcomes in the published studies. Estroff et al. (1992) found in 17 cases with DWV: four with mild ventriculomegaly, eight with noncentral nervous system anomalies; five had an abnormal karyotype, six died in utero or during the neonatal period, two were severely handicapped and the other nine were developing normally. After 8 years, (Ecker et al., 2000) found different results when studying a larger group of 49 fetuses with DWV: 85% had other sonographically identifiable anomalies, the most common being ventriculomegaly and cardiac defects; 36% of available karyotypes were abnormal. Totally 13 infants with DWV survived the neonatal period and 7 of 13 were reported initially as normal infants, including six with an isolated finding of DWV. In the most recent article on prenatal detection of posterior fossa malformations, Limperopoulos et al. (2008) abandoned the use of the term DWV, but surprisingly now only have two different diagnoses for patients with vermian anomalies: either DWM or inferior vermian hypoplasia; vermian agenesis is not mentioned. Prenat Diagn 2009; 29: 372–380. DOI: 10.1002/pd

374

G. MALINGER ET AL.

Table 1—Definitions of cerebellar pathologies amenable to prenatal diagnosis Dandy Walker malformation Cystic dilatation of the fourth ventricle with enlarged fourth ventricle, hypoplastic vermis and elevation of the tentorium. Vermian agenesis In complete vermian agenesis, the vermis is absent. In partial vermian agenesis, part of the vermis is absent and the remaining part is anatomically of normal volume. Vermian hypoplasia Normally formed vermis but of smaller size. Rhombencephalosynapsis Fusion of the cerebellar hemispheres with different degrees of vermian agenesis. Cerebellar hypoplasia Normally formed cerebellum with a small transcerebellar diameter. Pontocerebellar hypoplasia Small cerebellum associated with a flat pons. Cerebellar atrophy Normally formed cerebellum that stops growing in the last trimester. Cannot be distinguished from hypoplasia unless images from early pregnancy showed a normal cerebellar size. Mega cisterna magna Enlarged cisterna magna (>10 mm) with otherwise normal cerebellar structures. Posterior fossa arachnoid cyst Retrocerebellar pocket of cerebrospinal fluid that does not communicate with the cisterna magna Blake pouch cyst (delayed closure of the vermis) Apparent communication between the fourth ventricle and the posterior fossa with a normal vermis; the vermian fissures and fastigium are also normal. Unilateral cerebellar damage Total or partial cerebellar destruction due to an prenatal insult (hemorrhage, infarction, infection).

They diagnosed inferior vermian hypoplasia when caudal growth of the inferior vermis over the fourth ventricle remained incomplete after 18–20 weeks of gestation. The outcome of inferior vermis hypoplasia detected by prenatal magnetic resonance imaging (MRI) in 19 fetuses was found to be good; only three infants (with confirmed postnatal diagnosis) demonstrated motor and language delays and functional difficulties, six of the infants had normal postnatal MRI studies and all had normal development (Limperopoulos et al., 2006). We dare to suggest that at least some of these fetuses had delayed rotation or closure of the vermis and not true inferior hypoplasia and are similar to the cases described by Zalel et al. (2006). We suggest that the definition of inferior vermis hypoplasia should not be made based only on morphological criteria, but should be supplemented by abnormal vermian biometry. This concept is reinforced by the fact that also in children and adults, isolated inferior vermian hypoplasia is not considered, by some authors, to represent an abnormal condition (Patel and Barkovich, 2002). The terms hypoplasia and agenesis are used interchangeably in the literature on vermian abnormalities but actually there are completely different definitions of these entities as clearly explained by Guibaud and des Portes (2006): ‘Agenesis’ is defined, as either complete or partial absence of an anatomical structure. In complete vermian agenesis, the vermis is absent, whereas in partial vermian agenesis, part of the vermis is absent and the remaining part is anatomically of normal volume. Due to the craniocaudal development of the vermis, partial agenesis involves its inferior (caudal) part. In contrast, hypoplasia means: ‘a small but complete anatomical structure with a congenital volume diminution’. Copyright  2009 John Wiley & Sons, Ltd.

Therefore, the term inferior vermis hypoplasia used by Limperopoulos et al. (2006, 2008) should actually be inferior vermis agenesis. When a suspicion of an abnormal vermis is raised, it is crucial to be familiar with and to attempt to differentiate between the different entities that may be associated with this anomaly: DWM, vermian hypoplasia, vermian agenesis and vermian agenesis with associated brainstem malformations (MTRS and cobblestone lissencephaly). Using the multiplanar approach previously described (Figure 1), it is possible in most, but not in all, cases to differentiate between the entities (Table 2): 1. when all the measurements are according to the norms for gestational age, two vermian fissures are visualized with a ratio of 1 : 2 between the superior and inferior part, and a normal pons is seen but there is a communication between the fourth ventricle and the cisterna magna, delayed closure of the vermis should be diagnosed and a good prognosis is expected; 2. cerebellar pathologies may be associated with chromosomal aberrations (Ecker et al, 200) and represent an indication for amniocentesis; 3. when the vermis or part of it is small but all lobules are present and there are no other associated anomalies, vermian hypoplasia should be diagnosed. The prognosis is not yet clear since it can be a dominant benign trait or part of a genetic syndrome. Posterior vermian hypoplasia has been described in fragile X (Guerreiro et al., 1998; Mostofsky et al., 1998), so in these cases we recommend testing the mother; 4. when the vermis or its inferior part is missing, vermian or partial vermian agenesis should be diagnosed Prenat Diagn 2009; 29: 372–380. DOI: 10.1002/pd

375

THE FETAL CEREBELLUM

Table 2—Imaging findings in patients with congenital cerebellar malformations Diagnosis

TCD

Vermis biometry

Sup/Inf ratio

Fastigium

Fourth ventricle

Pons

Prognosis

Delayed ‘closure’ Vermian hypoplasia Vermian agenesis DWM MTRS Pontocerebellar hypoplasia Rhombencephalosynapsis

N N N/S N N/S S S

N Small Small —/Small —/Small Small —

N N Abn Abn Abn N —

N N Abn Abn Abn N Abn

Open Open/closed Abn Abn Abn N Abn

N N N N ? Abn N

Good Variable Malformations? Poor Poor Poor Poor

TCD, Transverse cerebellar diameter; Sup/Inf, ratio between the superior and inferior portions of the vermis; N, normal; S, small; Abn, abnormal; Malformations?, depends on the presence of associated malformations; −, agenesis.

5.

6.

7.

8.

and the prognosis depends on the existence of associated malformations; when vermian agenesis is associated with enlarged cisterna magna, cystic dilatation of the fourth ventricle and elevation of the tentorium, DWM should be diagnosed and the prognosis is usually poor but when there are no other anomalies and the vermis is actually hypoplastic, the prognosis may be better (Klein et al., 2003); when the fourth ventricle is enlarged with an abnormal fastigium, and the vermis is not clearly observed in the coronal plane, molar tooth features should be searched in the axial plane, including the interpeduncular fossa, cerebellar peduncles and brain stem in order to diagnose molar tooth related syndromes which are usually associated with mental retardation; when vermis hypoplasia is associated with a small transverse cerebellar diameter (TCD) and the bulge of the pons is missing, pontocerebellar hypoplasia should be diagnosed and the prognosis is universally grim; when the vermis is not present, the TCD is small and the cerebellar folia are continuous from one hemisphere to the other, rhombencephalosynapsis should be diagnosed and the prognosis is reserved in the majority of the cases. PREMATURE DIAGNOSIS OF ABNORMAL VERMIAN FORMATION

The current understanding of cerebellar formation seems to support the theory that the cerebellar vermis should finish its development at around 18 weeks, but this gestational landmark has never been fully proven. Bromley et al. (1994) studied 897 fetuses between 13 and 21 weeks of pregnancy and found that at 16 weeks 13% of the fetuses showed an ‘open vermis’ (an apparent communication between the fourth ventricle and the cisterna magna), at 17 weeks only 6% have a similar finding and after 18 weeks all the fetuses were found to have a ‘closed vermis’. The main limitation of this pivotal study was that only 77 fetuses were evaluated after 18 weeks. The authors concluded that the diagnosis of what they called DWV should not be made before 18 weeks. Babcook et al. (1996) found similar results in a study of 19 formalin-fixed fetuses at different gestational ages, between 11 weeks 5 days and 21 weeks, by Copyright  2009 John Wiley & Sons, Ltd.

the use of MRI and US, and recommended caution in the diagnosis at least before 18 weeks of pregnancy. Pilu et al. (2000) approached this issue when commenting on the results of the study by Carroll et al. (2000) and also mentioned the study by Ecker et al. (2000) that found a normal outcome in 7 of 13 infants with prenatally diagnosed DWV. According to Pilu et al. (2000), the diagnosis of an abnormal cerebellar vermis may be not possible in all cases. An early diagnosis of abnormal vermian formation is usually suspected during examinations performed before 20 weeks of pregnancy, but we have reviewed cases with similar findings as late as 27 weeks of pregnancy. These patients usually present with the description of an enlarged fourth ventricle and/or a ‘communication’ between the fourth ventricle and the cisterna magna. Characteristically these ‘findings’ are isolated and the size of the cisterna magna is normal. This phenomenon may be due in some cases, studied during the late second trimester and third trimester, to a very common artifact caused when increasing the angle from the transthalamic axial plane more than 15◦ in order to obtain the transcerebellar plane. The resulting images are actually semi to full coronal planes and therefore may show either a large cisterna magna, or an artefactual communication between the fourth ventricle and the cisterna magna (Laing et al., 1994). Bronshtein et al. (1998) described an apparently enlarged fourth ventricle in a series of 21 fetuses studied between 14 and 16 weeks of gestation; follow-up of 19 of them until 1 year was normal. Later on, Zalel et al. (2006) reported seven patients, six of them studied before 24 weeks of gestation, referred for evaluation because of suspected large cisterna magna and possible vermian agenesis; using US sagittal planes, complemented in some cases by MRI, the authors found that in these cases the vermis was apparently normal but superiorly rotated. We have observed similar cases in fetuses with an otherwise normal anatomy and found after repeated examinations that in most of the cases the ‘communication’ between the fourth ventricle and the cisterna magna disappears as the vermis enlarges (Figure 2). This may either be the result of normal variations in the development of the cerebellum; or increased pressure produced by the developing vermis may cause invagination and dehiscence of a Blake’s pouch cyst; or as postulated by Robinson and Goldstein (2007), the late Prenat Diagn 2009; 29: 372–380. DOI: 10.1002/pd

376

G. MALINGER ET AL.

Figure 2—Longitudinal US examinations in a fetus with suspected vermian anomaly. (A) Axial plane at 15 weeks showing a large cystic fourth ventricle (arrow) in communication with the cisterna magna. Note the widely separated cerebellar hemispheres (c); (B) at 19 weeks the fourth ventricle remains prominent (arrow); (C) midsagittal plane at 24 weeks shows a fully developed vermis with a normal sized fourth ventricle (arrow) and no communication with the cisterna magna; the vermian fissures are still not visible; (D) coronal plane at 24 weeks shows the hypoechogenic cerebellar hemispheres and the hyperechogenic vermis (arrows); (E) sagittal plane at 28 weeks shows the fully developed vermis with primary (1) and secondary (2) fissures and normal shaped fourth ventricle (3). Note the bulging of the pons (4)

closure of the fourth ventricle may be caused by a late fenestration of Blake’s pouch. In some cases, the vermis does not ‘close’ in utero and even though it is morphologically normal, an erroneous diagnosis of inferior vermis hypoplasia may be made. This observation may explain the finding by Limperopoulos et al. (2006) that in 32% of fetuses diagnosed with inferior Copyright  2009 John Wiley & Sons, Ltd.

vermian hypoplasia in utero, the postnatal MRI was normal. Presently, we believe that the diagnosis of the different forms of vermian hypoplasia should not be performed before 24 weeks of gestation. By this time, the neurosonographic examination of the vermis should enable accurate measurements of the vermian diameters and Prenat Diagn 2009; 29: 372–380. DOI: 10.1002/pd

THE FETAL CEREBELLUM

377

surface in the midsagittal plane and demonstration of a normal sized fourth ventricle with a triangular shape in midsagittal planes and the primary and sometimes also the secondary fissure originating from the fastigium. In cases in which the vermis remains open and there are no associated malformations, one should be extremely cautious in making the diagnosis of inferior vermis hypoplasia and the vermis should be measured and compared to norms (Malinger et al., 2001).

DIFFICULTIES IN THE ULTRASONOGRAPHIC DIFFERENTIATION BETWEEN THE CEREBELLAR HEMISPHERES AND THE VERMIS The fetal vermis is characterized by US as a distinct structure more echogenic than the adjacent cerebellar hemispheres. Since the vermis and cerebellar hemispheres have quite similar structures, it is not clear why they demonstrate different echogenicities. A possible answer to this question resides in the difference between the cerebellar folia and the vermian lobules: the cerebellar folia are shallow and the cerebrospinal fluid (CSF) and arachnoid vessels are located superficially; on the other hand, the vermian fissures reach deeply into the vermis, enabling a deeper penetration of CSF and arachnoid vessels. The interface between vermis, arachnoid vessels and CSF produces the characteristic hyperechogenic appearance of the vermis. When the vermis is totally or partially absent and there is no cystic formation protruding from the fourth ventricle, as occurs in MTRS, both cerebellar hemispheres remain in close contact. In cases of complete vermian agenesis, the amount of CSF and arachnoid vessels in the zone of contact is increased producing a hyperechogenic pattern that may be erroneously interpreted as a normal vermis (Figure 3). In cases of inferior vermian agenesis, the midline may be filled with the inferior part of

Figure 3—Axial plane at 25 weeks of gestation shows an apparently normal cerebellum in a child diagnosed postnatally as suffering from vermian agenesis. Note that the cerebellar hemispheres do not have a normal round shape and the echogenic midline structure does not have the appearance of the normal vermis (arrow) Copyright  2009 John Wiley & Sons, Ltd.

Figure 4—High magnification sagittal plane at 33 weeks of gestation in a fetus with vermian agenesis shows the cerebellar hemisphere displaced towards the midline. The arrow marks the border between the superior portion of the vermis and the cerebellum displaced towards the midline. Note the abnormal shape of the fourth ventricle (4) and absence of primary and secondary fissures. Compare with Figures 1C and 2E. cisterna magna (CM)

the cerebellum and both structures may mimic a normal vermis (Figure 4). Isolated vermian agenesis may be extremely difficult to diagnose even when there is a family history (Figure 5). Anomalies usually associated with vermian agenesis, such as occipital encephalocele, cystic kidneys or polydactyly should direct the ultrasonographer to a thorough examination of the vermis and may enable the correct diagnosis. In cases of suspected isolated vermian agenesis without a clear communication between the fourth ventricle and the cisterna magna, we recommend to search for indirect signs that may help in the diagnosis. An effort should be made to visualize the vermis in true orthogonal planes. In the axial plane, a smaller than expected transverse cerebellar diameter may be the first clue in the diagnosis of rhombencephalosynapsis, a rare congenital condition characterized by partial or total agenesis of the vermis with fusion of the cerebellar hemispheres; in these patients, the folia is usually continuous between the hemispheres. Another clue is that the cerebellar hemispheres do not protrude into the cisterna magna and appear continuous in the axial plane (Figure 3). The midsagittal plane enables visualization of the vermis, fourth ventricle, pons, medulla and cisterna magna and the relationship between them. In this plane, the vermis may be extremely difficult to differentiate from an impinging cerebellum, but in these cases the shape of the fourth ventricle should be abnormal (Figure 4) and the vermian fissures difficult to visualize. The operator should look for other possible clues including the presence of a thin pons and medulla as seen in pontocerebellar hypoplasia. In suspected cases, we found the study of the coronal plane useful in order to differentiate between the cerebellar folia and the vermian fissures (Figures 2D, 5B). Although the general opinion is that fetal MRI has better diagnostic capabilities than US, our experience is similar to that of Limperopoulos et al. (2008), that both Prenat Diagn 2009; 29: 372–380. DOI: 10.1002/pd

378

G. MALINGER ET AL.

Figure 5—Fetus at 22 weeks of gestation with a previous sibling known as suffering from Joubert’s syndrome. Axial (A), coronal (B) and sagittal (C) planes. The vermis is not clearly observed in the axial plane but seems to be, at least, partially present in the coronal plane. The sagittal plane is also not conclusive; the arrow may indicate either the primary fissure or the end of an abnormally small vermis with an impinging cerebellar hemisphere filling the midline. The pons and brain stem appear normal. Follow-up US and MRI were performed and found to be normal

false negative and false positive diagnoses of vermian anomalies may be made by the use of this technique. These facts are important particularly for their medicolegal implications and both the physician and families should be aware of the current limitations of fetal imaging in the diagnosis of isolated vermian pathologies. LATE DEVELOPMENT OF CEREBELLAR HYPOPLASIA/ATROPHY From the imaging point of view, neocerebellar diseases may result in the development of cerebellar hypoplasia or atrophy. Cerebellar hypoplasia is characterized by an apparently morphologically normal but small cerebellum with a normal or small sized cisterna magna (Figure 6). Cerebellar atrophy is defined as a cerebellum that was initially normal with a progressive increase in the size of the fissures in comparison to the size of the foliae (Poretti et al., 2008). Although the definition clearly separates between these two entities, in clinical practice, the distinction is much less obvious; this fact has produced confusion in the existing literature. While most cases of cerebellar atrophy start after birth, there are a group of rare disorders with cerebellar hypoplasia that have an onset during fetal or early neonatal life—the pontocerebellar hypoplasias (PCH). Even in these disorders, the hypoplasia is combined with progressive cerebellar degeneration that continues after birth. Cases amenable to prenatal diagnosis are rare; and only isolated cases and small series have been published (Mitra et al., 1999; Patel et al., 2006; Phadke et al., Copyright  2009 John Wiley & Sons, Ltd.

Figure 6—Cerebellar hypoplasia at 23 weeks of gestation in an otherwise normal fetus. The measured transcerebellar diameter (19.7 mm) was below the second percentile for gestational age. The vermis was apparently normal

2007). These cases are usually associated with severe central nervous system (CNS) and/or non-CNS anomalies, and/or chromosomal disorders (Smith et al., 2005). In most of these patients, the associated findings led to the depiction of the abnormal cerebellum. Steinlin et al. (2007) reported on 24 children with PCH type 2, characterized clinically by the findings of accompanying progressive microcephaly and dyskinetic movement disorder. Fourteen of these patients underwent a second trimester US that was considered normal and the head circumference of all 24 children was normal at birth, but later on they developed microcephaly. This study highlights the main problem in prenatal diagnosis of PCH, in Prenat Diagn 2009; 29: 372–380. DOI: 10.1002/pd

379

THE FETAL CEREBELLUM

these patients the only chance for diagnosis of the disease remains upon the depiction of a small cerebellum during the late stages of pregnancy at the time of routine third trimester examinations for fetal weight estimation. Since the evaluation of the brain is not performed during this examination, the probability of making a correct diagnosis is low. Patients with a family history of cerebellar hypoplasia and perhaps those with borderline TCD measurements should be evaluated at regular intervals during pregnancy with particular attention to the development not only of the cerebellum but also of the ventral pons. The family should be informed that even in cases with an apparently normal developing cerebellum and pons prenatal diagnosis of these conditions may not be always possible due to different possible phenotypes. DIFFERENTIAL DIAGNOSIS OF UNILATERAL CEREBELLAR FINDINGS Unilateral cerebellar findings of different echogenicities may be observed at any time of pregnancy and are usually the result of a focal insult to the cerebellum (Figure 7). These insults may be caused by hemorrhage

or infarction of the cerebellar tissue or the surrounding structures or may be the end result of intrauterine infection. The main diagnostic problem resides in the fact that the insult may occur at any time during pregnancy and thus may escape prenatal diagnosis. Fetal cerebellar hemorrhage is rare but a number of cases have recently been reported (Ranzini et al., 1998; Gorincour et al., 2006; Malinger et al., 2006). The initial lesion is hyperechogenic in cases due to hemorrhage or hemorrhagic infarction. Follow-up examinations usually show gradual reduction in the size of the affected hemisphere with or without a cystic lesion (Ranzini et al., 1998; Malinger et al., 2006). In rare circumstances, the hemorrhage may involve the contralateral cerebellar hemisphere and also the pons and midbrain (Lerner et al., 2006). The prognosis in these cases is difficult to predict but at least in some reported cases and also according to our personal experience these children may develop normally if the lesion does not involve the vermis. Cytomegalovirus infection may affect the cerebellum in different ways, cerebellar calcifications, cerebellar and/or vermian dysgenesis and cerebellar cysts have been reported (Malinger et al., 2003). The differential diagnosis in patients with hyperechogenic intracerebellar findings should include a possible nonhemorrhagic vascular malformation. In these cases, the lesion will remain static during pregnancy; T1 weighted MR images may be useful to rule out the presence of hemorrhage (Guibaud et al., 2003). In some cases, it may be difficult to differentiate between a cerebellar lesion and dural sinus thrombosis adjacent to the cerebellum (Laurichesse Delmas et al., 2008); the use of MRI has been also proposed in this case as an aid in the differential diagnosis. To obtain an accurate diagnosis in these cases is important since usually they carry a good prognosis.

CONCLUSION The introduction of advanced imaging techniques has greatly improved our diagnostic capabilities regarding the anatomy and pathology of the fetal cerebellum. In spite of these advances, the evaluation of the fetal cerebellum remains problematic. Longitudinal examinations and referral to specialized referral centers should be offered in cases with apparently isolated cerebellar conditions, when the anomaly is not clearly defined or when it is suspected early in pregnancy. The parents should be informed that a definitive prenatal diagnosis is not always possible.

REFERENCES Figure 7—Cerebellar hemorrhage at 18 (A) and 21 (B) weeks showing almost complete hemispheric atrophy. Termination of pregnancy was performed at 23 weeks; the autopsy failed to demonstrate the existence of the affected cerebellar hemisphere and vermis Copyright  2009 John Wiley & Sons, Ltd.

Adamsbaum C, Moutard ML, Andr´e C, et al. 2005. MRI of the fetal posterior fossa. Pediatr Radiol 35: 124–140. Babcook CJ, Chong BW, Salamat MS, Ellis WG, Goldstein RB. 1996. Sonographic anatomy of the developing cerebellum: normal Prenat Diagn 2009; 29: 372–380. DOI: 10.1002/pd

380

G. MALINGER ET AL.

embryology can resemble pathology. AJR Am J Roentgenol 166: 427–433. Blake JA. 1900. The roof and lateral recesses of the fourth ventricle, considered morphologically and embryologically. J Comp Neurol 10: 79–108. Bromley B, Nadel AS, Pauker S, Estroff JA, Benacerraf BR. 1994. Closure of the cerebellar vermis: evaluation with secod trimester US. Radiology 1993: 761–763. Bronshtein M, Zimmer EZ, Blazer S. 1998. Isolated large fourth ventricle in early pregnancy–a possible benign transient phenomenon. Prenat Diagn 18: 997–1000. Carroll SG, Porter H, Abdel-Fattah S, Kyle PM, Soothil PW. 2000. Correlation of prenatal ultrasound diagnosis and pathologic findings in fetal brain abnormalities. Ultrasound Obstet Gynecol 16: 149–153. Ecker JL, Shipp TD, Bromley B, Benacerraf B. 2000. The sonographic diagnosis of Dandy-Walker and Dandy Walker variant: associated findings and outcomes. Prenat Diagn 20: 328–332. Estroff JA, Scott MR, Benacerraf BR. 1992. Dandy-Walker variant: prenatal sonographic features and clinical outcome. Radiology 185: 755–758. Gorincour G, Rypens F, Lapierre C, Costa T, Audibert F, Robitaille Y. 2006. Fetal magnetic resonance imaging in the prenatal diagnosis of cerebellar hemorrhage. Ultrasound Obstet Gynecol 27: 78–80. Guerreiro MM, Camargo EE, Kato M, et al. 1998. Fragile X syndrome. Clinical, electroencephalographic and neuroimaging characteristics. Arq Neuropsiquiatr 56: 18–23. Guibaud L, des Portes V. 2006. Plea for an anatomical approach to abnormalities of the posterior fossa in prenatal diagnosis. Ultrasound Obstet Gynecol 27: 477–481. Guibaud L, Garel C, Annie B, et al. 2003. Prenatal diagnosis of capillary telangiectasia of the cerebellum—ultrasound and MRI features. Prenat Diagn 23: 791–796. Halawa SM, Bar-Hava I, Lev D, Vinkler H, Lerman-Sagie T, Malinger G. 2007. Prenatal diagnosis of cerebellar anomlies by dedicated neurosonography. Ultrasound Obstet Gynecol 30: 389. ISUOG guidelines. 2007. Sonographic examination of the fetal central nervous system: guidelines for performing the basic examination and the fetal neurosonogram. Ultrasound Obstet Gynecol 29: 109–116. Klein O, Pierre-Kahn A, Boddaert N, Parisot D, Brunelle F. 2003. Dandy-Walker malformation: Pre-natal diagnosis and prognosis. Childs Nerv Syst 19: 484–489. Laing FC, Frates MC, Brown DL, Benson CB, Di Salvo DN, Doubilet PM. 1994. Sonography of the fetal posterior fossa: false appearance of mega-cisterna magna and Dandy-Walker variant. Radiology 192: 247–251. Laurichesse Delmas H, Winer N, Gallot D, et al. 2008. Prenatal diagnosis of thrombosis of the dural sinuses: report of six cases, review of the literature and suggested management. Ultrasound Obstet Gynecol 32: 188–198. Lerner A, Gilboa Y, Gerad L, Malinger G, Kidron D, Achiron R. 2006. Sonographic detection of fetal cerebellar cavernous hemangioma with in-utero hemorrhage leading to cerebellar hemihypoplasia. Ultrasound Obstet Gynecol 28: 968–971. Limperopoulos C, Robertson RL, Estroff JA, et al. 2006. Diagnosis of inferior vermian hypoplasia by fetal magnetic resonance imaging: potential pitfalls and neurodevelopmental outcome. Am J Obstet Gynecol 194: 1070–1076. Limperopoulos C, Robertson RL Jr, Khwaja OS, et al. 2008. How accurately does current fetal imaging identify posterior fossa anomalies? AJR Am J Radiol 190: 1637–1647.

Copyright  2009 John Wiley & Sons, Ltd.

Malinger G, Ginath S, Lerman-Sagie T, Watemberg N, Lev D, Glezerman M. 2001. The fetal cerebellar vermis: normal development as shown by transvaginal ultrasound. Prenat Diagn 21: 687–692. Malinger G, Lev D, Zahalka N, et al. 2003. Fetal cytomegalovirus infection of the brain: the spectrum of sonographic findings. AJNR Am J Neuroradiol 24: 28–32. Malinger G, Zahalka N, Kidron D, Ben-Sira L, Lev D, Lerman-Sagie T. 2006. Fatal outcome following fetal cerebellar haemorrhage associated with placental thrombosis. Eur J Paediatr Neurol 10: 93–96. Mitra AG, Salvino AR, Spence JE. 1999. Prenatal diagnosis of fatal infantile olivopontocerebellar hypoplasia syndrome. Prenat Diagn 19: 375–378. Mostofsky SH, Mazzocco MM, Aakalu G, Warsofsky IS, Denckla MB, Reiss AL. 1998. Decreased cerebellar posterior vermis size in fragile X syndrome: correlation with neurocognitive performance. Neurology 50: 121–130. Parisi MA, Dobyns WB. 2003. Human malformations of the midbrain and hindbrain: Review and proposed classification scheme. Mol Genet Metab 80: 36–53. Patel S, Barkovich AJ. 2002. Analysis and classification of cerebellar malformations. AJNR Am J Neuroradiol 23: 1074–1087. Patel MS, Becker LE, Toi A, Armstrong DL, Chitayat D. 2006. Severe, fetal-onset form of olivopontocerebellar hypoplasia in three sibs: PCH type 5? Am J Med Genet A 140: 594–603. Phadke S, Puri R, Phadke R. 2007. Severe form of congenital cerebral and cerebellar atrophy: a neurodegenerative disorder of fetal onset. J Clin Ultrasound 35: 347–350. Phillips JJ, Mahony BS, Siebert JR, Lalani T, Fligner CL, Kapur RP. 2006. Dandy-Walker malformation complex: correlation between ultrasonographic diagnosis and postmortem neuropathology. Obstet Gynecol 107: 685–693. Pilu G, Romero R, De Palma L, Jeanty P, Burdine C, Hobbins JC. 1987. Ultrasound investigation of the posterior fossa in the fetus. Am J Perinatol 4: 155–159. Pilu G, Visentin A, Valeri B. 2000. The Dandy-Walker complex and fetal sonography. Ultrasound Obstet Gynecol 16: 115–117. Poretti A, Wolf NI, Boltshauser E. 2008. Differential diagnosis of cerebellar atrophy in childhood. Eur J Paediatr Neurol 12: 155–167. Ranzini AC, Shen-Schwarz S, Guzman ER, Fisher AJ, White M, Vintzileos AM. 1998. Prenatal sonographic appearance of hemorrhagic cerebellar infarction. J Ultrasound Med 17: 725–727. Robinson AJ, Goldstein R. 2007. The cisterna magna septa: vestigial remnants of Blake’s pouch and a potential new marker for normal development of the rhombencephalon. J Ultrasound Med 26: 83–95. Smith AS, Levine D, Barnes PD, Robertson RL. 2005. Magnetic resonance imaging of the kinked fetal brain stem: a sign of severe dysgenesis. J Ultrasound Med 24: 1697–1709. Smith PA, Johansson D, Tzannatos C, Campbell S. 1986. Prenatal measurement of the fetal cerebellum and cisterna cerebellomedullaris by ultrasound. Prenat Diagn 6: 133–141. Steinlin M, Klein A, Haas-Lude K, et al. 2007. Pontocerebellar hypoplasia type 2: variability in clinical and imaging findings. Eur J Paediatr 11: 146–152. Zalel Y, Gilboa Y, Gabis L, et al. 2006. Rotation of the vermis as a cause of enlarged cisterna magna on prenatal imaging. Ultrasound Obstet Gynecol 27: 490–493.

Prenat Diagn 2009; 29: 372–380. DOI: 10.1002/pd