Inferior vermian hypoplasia – preconception, misconception

There is considerable confusion in the literature regarding the terminology used when describing abnormalities of the cerebellum and of the vermis in particular. Terminology such as ‘closure of the fourth ventricle’, ‘craniocaudal growth of the vermis’ and ‘inferior vermian hypoplasia’, as well as the numbering (using Roman numerals) of the cerebellar lobules from anterior to posterior, has left us with the preconception and misconception that the vermis grows from superior to inferior, and that partial agenesis or hypoplasia always involves the inferior lobules. In light of recent advances in our understanding of the embryology of the cerebellum and cisterna magna, certain terminology and concepts can be demonstrated to be incorrect and should be abandoned. This Editorial reviews the normal development of the cerebellum, describing and dispelling several misconceptions regarding both normal and abnormal cerebellar development, with specific reference to the cerebellar vermis. Examples of cerebellar and vermian anatomy at pathology, in-vitro and in-vivo fetal magnetic resonance imaging (MRI) and pre- and postnatal imaging are reviewed and correlated with cerebellar embryology, phylogeny, somatotopic mapping and functional MRI. The evidence indicates that the cerebellar vermis develops more in a ventral to dorsal direction than in a superior to inferior one and, therefore, that the concept of ‘inferior vermian hypoplasia’ is incorrect. Three possible categories of vermian anomaly are seen: it may not necessarily be the inferior vermis that is hypoplastic; it may not only be the inferior vermis that is hypoplastic; or it may not be vermian hypoplasia at all. The term ‘inferior vermian hypoplasia’ should only be used if it can be proved that only the inferior vermis is abnormal. There is no generic term which encompasses all the various etiologies that can cause a small vermis; thus, more appropriate terminology may be ‘vermian hypoplasia’ or ‘vermian dysplasia’, with ‘neovermian hypoplasia’ in cases in which the central lobules are proved to be abnormal. The neural tube develops, during the 3rd and 4th weeks of embryogenesis, as a longitudinal groove along the dorsal layer of the trilaminar germ cell disc1. This groove deepens until its edges meet and fuse over the top, forming a hollow tube, which closes off at the cranial and caudal ends on days 25 and 28, respectively. The cranial end of the neural tube differentiates into three distinct regions: the prosencephalon, caudal to this the mesencephalon, and even more caudally the rhombencephalon, each enclosing a brain vesicle containing amniotic fluid initially. The rhombencephalon segments into eight rhombomeres. The future cerebellum originates in the alar plate of the most rostral adjacent pair of these rhombomeres; namely, rhombomeres 1 and 22. Several genes (Otx, Gbx, FGF, Hox)3 play a role in the formation and function of the isthmic organizer at the junction of the mesencephalon and rhombomere 1, which in turn regulates cerebellar development through release of hormonal factors4-6. The more caudal rhombomere 2 gives rise to the germinal matrix of the ventricular zone, which forms the deep cerebellar nuclei (fastigial, globose, emboliform and dentate from medial to lateral) and the Purkinje cell layer. The more rostral rhombomere 1 gives rise to the germinal matrix of the rhombic lips. Initially, this forms the external granular layer of the cerebellum, which subsequently migrates deeper to form the internal granular layer, a process which is completed by approximately 2 years postnatally. The cerebellar vermis per se is not formed through fusion of the adjacent developing cerebellar hemispheres but develops as a direct proliferation of the mesial primordium, starting in the 5th week of gestation7, 8. Experimental evidence shows us that granule cells arising from the lateral upper rhombic lip migrate medially into the posterior cerebellum, whereas granule cells arising in the medial upper rhombic lip are confined to an anterior cerebellar distribution (Figure S1)8. Therefore, as the primordia are separate, the development of the posterior vermis is not dependent on that of the anterior vermis. This is evident in the case of rhombencephalosynapsis, in which the posterior vermis and most inferior lobule (nodulus (X)) can be preserved in the absence of the anterior lobe of the vermis9, 10. Thus, vermian hypoplasia can be segmental, and normal development of the inferior vermis is not dependent on normal development of the superior vermis. Key point: The vermis does not form through fusion of the cerebellar hemispheres, it develops from its own mesial primordium. Additionally, different parts of the vermis develop from different parts of the mesial primordium. Segmental vermian abnormalities can therefore exist. At around 8–10 weeks' gestation, a focal dilatation of the neural tube is seen in the dorsal aspect of the developing hindbrain; this is the rhombencephalic vesicle, the predecessor to the fourth ventricle. At this level in the brainstem, which is known as the open medulla, the two alar laminae do not touch in the midline dorsally, and the gap is bridged by a layer of tela choroidea11 known as the area membranacea, which forms the roof of the rhombencephalic vesicle. The hindbrain develops a kink, known as the dorsal pontine flexure, which causes a transverse crease to form in the area membranacea, dividing it into a more rostral anterior membranous area and a more caudal posterior membranous area (Figure 1). The cerebellum develops from the rhombic lips at the cranial end of the area membranacea and the vermis and cerebellum grow exophytically, inferiorly and laterally, to cover it. Due to its thinness, this layer of tela choroidea has not, until recently, been resoluble by in-vivo imaging, thus giving the false impression that the fourth ventricle is open to the developing subarachnoid space initially and then closes due to caudal growth of the overlying vermis12. This apparent developmental process is therefore often referred to in older literature as ‘closure’ of the fourth ventricle. This is typically complete by around 18 weeks' gestation, although physiological variation may give the appearance that the vermis is incomplete at the time of the initial mid-trimester assessment13. As the vermis grows caudally it invaginates into the rhombencephalic vesicle, and the posterior membranous area protrudes beneath the vermis into the overlying meninx primitiva14, 15. This evagination, first described in 1900, is known as Blake's pouch11, and where Blake's pouch constricts to pass through the cerebellar vallecula (the normal space inferior to the vermis, superior to the nucleus gracilis and medial to the cerebellar hemispheres) it is known as Blake's metapore (Figure 1d). Even though Blake's pouch itself lies within the subarachnoid space of the developing cisterna magna, it is a direct extension of the fourth ventricle and therefore the fluid contained in Blake's pouch is intraventricular. Normal linear echoes (the cisterna magna septa), which are typically seen in the fetal and neonatal cisterna magna and are most often described as bridging arachnoid septations16, have recently been shown to represent the walls of Blake's pouch and the cisterna magna septa are a potential marker for normal development15 (Figure 2a). The future cisterna magna therefore forms in two compartments: a mesial compartment between the cisterna magna septa, which is derived from the rhombencephalic vesicle (Blake's pouch), and compartments lateral to the cisterna magna septa which develop through cavitation of the meninx primitiva overlying the surface of the brain, forming the subarachnoid space proper(Figures 2b and c). Blake's pouch usually, but not always, fenestrates to a variable degree11, 17 down to the obex (the inferior recess of the fourth ventricle), which leads to communication between the mesial ventricular-derived compartment and the true subarachnoid space of the cisterna magna. Fenestration and disappearance of Blake's pouch thus leaves an opening at Blake's metapore15, 17, 18, which is known as the foramen of Magendie, allowing communication between the fourth ventricle and cisterna magna. Thus, the foramen of Magendie does not demarcate the true junction between the ventricular system and subarachnoid space of the cisterna magna. A small communication beneath the vermis, between the fourth ventricle and the ‘cisterna magna’, often described in the literature on ultrasound12, 15, 19-21 and sometimes seen on mid-sagittal images, therefore represents the normal Blake's metapore. Imaging the posterior fossa in the semicoronal plane will show this normal opening12, 15, 21 (Figure 3), but can give a false appearance of a vermian defect22 and, unfortunately, is often described wrongly as ‘Dandy–Walker variant’ (vide infra). This error can be avoided by making sure that the cavum septi pellucidi is included in the image, thus ensuring that the scan plane is truly axial or modified-axial. Key point: A small gap between the vermis and the brainstem is normal and is known as Blake's metapore as it contains the neck of Blake's pouch. Once Blake's pouch fenestrates, the metapore is known as the foramen of Magendie. ‘Ontogeny recapitulates phylogeny’, or embryology repeats evolution. This important principle states that the development observed during embryology is like a ‘time-lapse photography’ rendition of the various steps taken during evolution; thus, structures that evolved first also develop first in the embryo, although the separate steps become somewhat merged. The oldest part of the cerebellum, the archicerebellum, comprises the bilateral flocculi and the mesial nodulus (X), and is known as the flocculonodular lobe (Figure 4a and S2)23. Functionally, this lobe plus some of the adjacent uvula (IX) comprise the vestibulocerebellum, which has connections with the vestibular nuclei (which, although situated within the brainstem, are considered surrogate deep cerebellar and from the superior and is in and in are all and are first seen in and are the to The part of the cerebellum to is the which comprises the lobules of the anterior lobe of the vermis and, although often but in by the the more caudal lobules of the posterior lobe of the vermis and some of the adjacent uvula The plus adjacent in the cerebellar hemispheres is known as the and, due to its connections with the and connections the deep cerebellar nuclei to the it is in and of and this is first seen in and, in is in and The part of the cerebellum to is the known as the or which comprises the most rostral lobules of the posterior vermis the and plus the of the cerebellar Functionally, it is with and and it is to and this is seen in only and it is in it is therefore the to both in and the most to the and thus is one of the most used for normal cerebellar development in the A of development is seen in the deep cerebellar which the older the are the most medial within the and with the archicerebellum, by the and emboliform which are more lateral and with the and the the most lateral which are with the therefore from an and than the anterior lobe developing first by the posterior the cerebellum develops with the most rostral and most caudal parts initially adjacent to each with the more recent structures subsequently developing between these older to the opening of a in which the are the first to and the Thus, the more anterior and posterior lobules and the medial deep cerebellar nuclei and the more central hemispheres and most lateral deep cerebellar nuclei Key point: development of the vermis more in a direction than in a evidence in of this of development is seen when at the most recently somatotopic of the cerebellum and This an initially in from superior to inferior, the posterior lobe with the anterior when at the of the anterior and posterior as a on of the cerebellar not only does it to it also that for the more or older of the and first and are in the to the brainstem and also to the and uvula whereas and such as the of the and lips and the of and are more and in the more central lobules of the vermis, and and and of the of the cerebellar hemispheres, respectively. Key point: mapping development of the vermis more in a direction than in a The cerebellar somatotopic has in been by functional of the cerebellum by (Figure the posterior lobe and anterior lobe as of each on of the cerebellar Key point: development of the vermis more in a direction than in a There is direct evidence that this of development is in the developing in The flocculonodular lobe is from the part of the cerebellum by the along with the that the anterior and posterior is the first to The to are the and (Figure There a of approximately weeks between is resoluble with in-vitro and also of approximately weeks between in-vitro with in-vivo At by fetal a area is seen the of each thus, at typically only lobules can be seen the lobules that are and only one these three lobules can be between the and (Figure the in can that this one develops into three separate the to which is the normal (Figure This may to of vermian hypoplasia at the and uvula (IX) can be as the and and with the lobules considered to have not Key point: both in and in development of the vermis more in a direction than in a In cases of ‘inferior vermian hypoplasia’ have no evidence that it is the inferior lobules that are often the inferior vermian lobules from the lobules of the vermis by of the vermian lobules do have on the and for the uvula (IX) and nodulus each have one the has two anterior and three posterior the has one and several and the has one and one Therefore, most of the lobules are A of the and which to on imaging, is that all three are to the by a whereas all the lobules have own (Figure for making such a from the lobules by giving separate is the that the and are in direct with and of the which form the of the cerebellar hemispheres (Figure Key point: that the different vermian lobules do not or are at the time of therefore, in ‘inferior vermian hypoplasia’ it may not be possible to that it is the inferior lobules that are in it may not be vermian hypoplasia at all. of the cerebellum also the for cerebellar of the vermian has been both and by has growth of both the superior and the inferior which grow on of the in with the inferior vermis linear growth of both not be seen and be a in the of the anterior and posterior Key point: development of the vermis more in a direction than in a described initially in with and to be the of of the of and that it is a more developmental the roof of the and to variable of vermian hypoplasia and variable of the fourth ventricular with variable (Figure The term ‘Dandy–Walker variant’ to with vermian hypoplasia but the of the of the and posterior 1). that the of with to the posterior fossa in the from two potential of vermian development it does not cover the fourth or of of the fourth ventricular to an Blake's pouch with and of the vermis. In these two are to Several have been described in of than the inferior vermian lobules typically all of the lobules and the of an of the lobules and seen the vermis, both the inferior lobe and the superior although the inferior lobules more the nodulus the uvula (IX) and in one case the with abnormal caudal of the germinal as if it by the of the fourth ventricle of Blake's (Figure to be also that the Purkinje cells and deep cerebellar nuclei from the germinal the the germinal matrix of the rhombic lips the confined to of rhombomere to normal is also an apparent of of in a superior to inferior due to and within the of the developing vermis due to from the isthmic Key point: may be that it is not only the inferior vermis that is abnormal in the to be a of of and the superior vermis is also abnormal. may for The term ‘inferior vermian hypoplasia’ is in as an to the ‘Dandy–Walker and has evolved as a of terminology such as ‘closure of the 4th ventricle’, ‘craniocaudal growth of the ‘craniocaudal and of the cerebellar hemispheres as well as the numbering of the cerebellar lobules from superior to are with the preconception that the vermis grows in a direction and due to development of the vermis, partial agenesis involves its inferior when development as it always 18 have a misconception that it be the inferior vermis that is even in a normal vermis is always an inferior gap in gestation, and the inferior lobules (X), uvula are the that to develop This has been shown on in-vitro imaging, when the pre- and therefore the lobules of flocculonodular (X), uvula can be seen as as weeks' gestation, whereas the and are not from each (Figure Thus, the inferior gap seen in this in be due to of the inferior it be due to the of the vermis, in the of development of the and when truly it may be due to a or in part of the vermis, the vermis will not as inferiorly as it when the is not is it that the are or is it that the have to grow and the into normal In of the appearance of the cerebellar it is more in cases of development, the flocculonodular lobe and lobules to the brainstem nodulus (X), uvula be and the lobules and cerebellar hemispheres and to in this the term ‘neovermian hypoplasia’ may be more A of this may be seen in in which in the lobules can be but the lobules are normal (Figure The may also if a the lobules which are developing at that allowing the lobules that to develop the lobules and adjacent cerebellum be but the and hemispheres be A of this may be seen in fetal when the lobules may be due to of the Purkinje cells which at the time of the but therefore, the lobules are but the lobules are normal (Figure Thus, on the and of the a different be and or a of and which have been shown to to the and Key point: vermian hypoplasia’ may be due to hypoplasia of vermian for the to in This may in certain cases of ‘inferior vermian hypoplasia’ may only The term inferior vermian hypoplasia’ has been used to that within ‘inferior vermian hypoplasia’ in which are no known or and this has a it is until recently, due to the of between ‘inferior vermian hypoplasia’ and Blake's pouch (vide in the literature these two of with have been included The lies in our even by imaging in to an Blake's pouch from true vermian both have an Blake's pouch and both have a vermis, and the in is on the appearance of the Thus, the is to that the vermis is to this to be sure that all of the vermis, and not the inferior for between Blake's pouch and vermian hypoplasia is the of the if it is in its normal on the inferior surface of the vermis of the this is with Blake's pouch it indicates that the anterior membranous area formed if it is on the inferior of the this indicates that the anterior membranous area formed (Figure of the vermis. This to is often to vermian hypoplasia is for or may be In one of and ‘inferior vermian hypoplasia’ abnormal even if the to be In postnatal imaging and normal in of cases of inferior vermian hypoplasia’, and the with postnatal of hypoplasia with only developmental in a of Key point: The of the may be to true vermian hypoplasia from Blake's pouch which have different In Blake's pouch (Figure is to be of both Blake's pouch and the of to of fluid into the subarachnoid space of the cisterna magna, with dilatation of the fourth the pouch with the fourth is a of communication between the pouch and the subarachnoid In most of the Blake's pouch is a normal in these the vermis grows even more caudally than it does in the mesial of Blake's metapore and dividing it into two lateral than in the of are also and therefore the normal of Blake's pouch does not In in with of of Blake's pouch causes it to and the vermis from the brainstem, this causes a gap between the inferior vermis and the brainstem, this can to the of ‘inferior vermian This of Blake's pouch has been of and in apparent of the posterior the and the vermis into a normal This is seen in and with Blake's pouch in is no vermian or to the ventricular system in a to normal appearance and of these the has of the vermis due to a Blake's pouch does not necessarily an 15, In one one of cases of Blake's pouch or cisterna magna in and of with no normal developmental at years the initial of vermian hypoplasia been In of cases of Blake's pouch seen in There two neonatal and eight one and the eight although seen in has also been in several cases in the literature that Blake's pouch can be if the of is by the of fetal or fetal which in dilatation and of the ‘cisterna of Blake's pouch contained within the cisterna through within the ventricular system of the in both the of and Blake's in the as can be demonstrated (Figure it is important to on the and of the to the developing brain itself can also from and cisterna magna may in represent Blake's pouch but with such that the vermis is not Key point: Blake's pouch can give the appearance of ‘inferior vermian hypoplasia’ when is no of the vermis at all. that or may cause this The vermis to grow in a thus giving rise to the preconception that when it is incomplete the the inferior vermis; evidence indicates that the vermis and cerebellum develop more in a with the more recent structures developing the older to the opening of a or in part of the vermis, or of Blake's pouch to of an normal vermis, can both give the appearance of ‘inferior vermian hypoplasia’ and can be to The term ‘inferior vermian hypoplasia’ is incorrect it is not necessarily the inferior vermis that is it may not only be the inferior vermis that is or the vermis may not be abnormal at all. The term should be in of vermian hypoplasia or vermian it can be proved that it is the inferior vermis that is and the that the may be due to or The part of the vermis on the and the of the which can structures at the time of the or which develop the this concept thus the of In certain cases it may even be more appropriate to the term ‘neovermian hypoplasia’ in of the that the and may be the only lobules like to the of for the in and of for for the in and of of for the in and and of for for the in and The is not for the or of by the than should be to the for the The may be in the of this with cerebellar hypoplasia and from and The more inferior parts of the cerebellum and vermis granule cells from the more lateral upper rhombic lip germinal whereas cells in the mesial germinal matrix are confined to an anterior cerebellar with from from of the The the flocculonodular lobe and adjacent uvula the the anterior lobe and the and medial hemispheres and the the and and the of the cerebellar from of the This with a development but with a The anterior lobe is a of the posterior lobe on of the cerebellar are at the cranial and caudal upper and that with from from and magnetic resonance imaging of the cerebellum a in which older from the than do This the somatotopic with from from of the cerebellum both the system and the numbering system for all of the vermian plus cerebellar lobules and is also known as the with from from and that all of the vermian lobules are and abnormal and the nodulus is and Normal for from of the nodulus Normal for the germinal matrix to have been inferiorly by the superior of Blake's pouch. Normal of the germinal matrix the in shows in the inferior vermis is also in the superior vermis and a of of in a superior to inferior Normal for with from from and In this at weeks' the and can be thus, the and uvula (IX) are even though is a gap the inferior vermis from the with from from In the lobules in are that are and during The and are from In fetal the lobules in are that are older and during embryogenesis, therefore Purkinje cells during to The lobules are from magnetic resonance (MRI) in a with The cerebellar vermis is but normal with normal and There is no ‘inferior vermian in the in the groove in the which also on the apparent in with germinal matrix in a different within Blake's pouch in with the fourth but the subarachnoid space of the cisterna magna has normal fluid in the within Blake's pouch and an and vermis giving a vermian hypoplasia