Assessment of Fetal Midbrain and Hindbrain in Mid-Sagittal Cranial Plane by Three-Dimensional Multiplanar Sonography

Ultrasound Obstet Gynecol 2014; 44: 581–587 Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/uog.13312

Assessment of fetal midbrain and hindbrain in mid-sagittal cranial plane by three-dimensional multiplanar sonography. Part 2: application of nomograms to fetuses with posterior fossa malformations

Z. LEIBOVITZ*†, C. SHKOLNIK‡, K. KRAJDEN HARATZ*, G. MALINGER§, I. SHAPIRO† and T. LERMAN-SAGIE*¶ *Unit of Fetal Neurology and Prenatal Diagnosis, Department of Obstetrics and Gynecology, Wolfson Medical Center, Holon, Israel, afﬁliated with the Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel; †Department of Obstetrics and Gynecology, Bnai Zion Medical Center, Haifa, Israel; ‡Carmel Medical Center, Haifa, Israel; §Ob-Gyn Ultrasound Unit, Lis Maternity Hospital, Tel Aviv Sourasky Medical Center and Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel; ¶Pediatric Neurology Unit, Wolfson Medical Center, Holon, afﬁliated with the Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

KEYWORDS: 3D ultrasound; CNS malformations; hindbrain; midbrain; nomogram; pons; posterior fossa malformations; vermis

ABSTRACT those with isolated enlarged posterior fossa cerebrospinal fluid spaces. Use of the nomograms enabled detection Objectives To apply fetal midbrain (MB) and hindbrain of an elongated tectum in fetuses with CMC, C-II and (HB) nomograms, developed using three-dimensional RES, and a flattened pontine belly in cases of CMC, PCH multiplanar sonographic reconstruction (3D-MPR) in the and VD. In the fetuses with VD, the nomograms enabled mid-sagittal cranial plane, to fetuses with known posterior division into three distinctive groups: (1) those with small fossa malformations. SIVD and APVD, (2) those with normal SIVD but small Methods In this retrospective study we examined sono- APVD, and (3) those with small SIVD but normal APVD. graphic volumes obtained by sagittal acquisition in 43 Conclusions Application of our new reference data, that fetuses diagnosed with posterior fossa abnormalities for the first time include the MB, enables accurate and evaluated in the mid-sagittal cranial plane, using diagnosis of brain malformations affecting the MB 3D-MPR, the following: MB parameters tectal length (TL) and HB and makes possible novel characterization of and anteroposterior midbrain diameter (APMD), and previously described features of posterior fossa anomalies. HB parameters anteroposterior pons diameter (APPD), Copyright © 2014 ISUOG. Published by John Wiley & superoinferior vermian diameter (SIVD) and anteropos- Sons Ltd. terior vermian diameter (APVD). Fetuses were grouped, according to malformation, into eight categories: cobblestone malformation complex (CMC, n = 3), Chiari-II INTRODUCTION malformation (C-II, n = 7), pontocerebellar hypoplasia (PCH, n = 2), rhombencephalosynapsis (RES, n = 4), Recent studies in neuroembryology, molecular genetics Dandy–Walker malformation (n = 8), vermian dysgen- and central nervous system (CNS) imaging have improved = = our understanding of disorders affecting development esis (VD, n 7), persistent Blake’s pouch cyst (n 6) 1 and megacisterna magna (n = 6).Ineachcaseandfor of the midbrain (MB) and hindbrain (HB) .The each subgroup, the MB-HB biometric parameters and mechanisms most investigated relate to anomalies of MB-HB formation and include abnormal anteroposterior their z-scores were evaluated with reference to our new 2 nomograms. mesencephalic-rhombencephalic patterning , disorders of the Reelin signaling pathway3, defective pial limiting Results The new MB-HB nomograms were able to membrane4 and disorders of the primary cilia5.MB identify the brainstem and vermian anomalies and and HB malformations are seldom isolated, frequently differentiate fetuses with MB-HB malformations from being associated with cerebral malformations6.Anew

Correspondence to: Dr Z. Leibovitz, Department of Obstetrics and Gynecology, Wolfson Medical Center, 62 Halohamim Street, POB 5, Holon 58100, Israel (e-mail: [email protected]; [email protected]) Accepted: 9 January 2014

Copyright © 2014 ISUOG. Published by John Wiley & Sons Ltd. ORIGINAL PAPER 582 Leibovitz et al. classification system has been proposed, based upon continuous configuration of the folia. DWM was defined embryology, genetics and neuroimaging7.Itisof by cystic dilatation of the fourth ventricle confluent with the utmost importance to diagnose accurately these an enlarged cisterna magna, elevation of the tentorium abnormalities, in order to give adequate counseling to the and a small dysplastic vermis. Regarding VD (previously parents. Using established nomograms of posterior fossa called Dandy–Walker variant), this term was used rather structures can facilitate the diagnosis. Many authors have than ‘vermian hypoplasia’ or ‘partial agenesis’ because, proposed nomograms for assessment of the fetal pons and even though the vermis was small, rotated slightly vermis8–11. However, despite the increased recognition upward and associated with a mid-sagittal connection of MB involvement in posterior fossa anomalies, until between the fourth ventricle and the cisterna magna, now there have been no reference data on the normal there was always both abnormal vermian lobulation and development of the mesencephalic structures in utero. fastigial shape; the tentorium was normally positioned. Inthefirstpartofthisstudy12, we established nomo- PBPC was defined as a cystic, thin-walled structure grams for assessment of the fetal MB and HB by protruding from the fourth ventricle and causing rotation three-dimensional multiplanar sonographic reconstruc- of a normal vermis without tentorial elevation or tion (3D-MPR) in the mid-sagittal cranial plane. In this distension of the cisterna magna. MCM was defined as second part of the study, we applied the nomograms to a symmetrically enlarged cisterna magna (anteroposterior fetuses with posterior fossa malformations in order to diameter >10 mm, normal cerebellum and brainstem and evaluate their accuracy, their ability to diagnose malfor- absence of vermian rotation and tentorial elevation). mations and their ability to differentiate between overt Using the 3D acquisitions retrieved from our archives, pathology and anomalies with more favorable outcome. we evaluated MB and HB measurements in the mid-sagittal cranial plane by 3D-MPR using 4D View software (GE Healthcare), as described in detail in Part 1 METHODS of the study12. MB measurements included: tectal length This research was conducted as a retrospective study and (TL) and anteroposterior midbrain diameter (APMD). HB was approved by the Institutional Review Board. Fetuses measurements included: anteroposterior pons diameter with known posterior fossa abnormalities were studied (APPD), superoinferior vermian diameter (SIVD) and with reference to our new MB-HB nomograms developed anteroposterior vermian diameter (APVD). These MB-HB by 3D-MPR12. biometric parameters and their z-scores were evaluated in Imaging data of the fetuses with CNS malformations each of the cases and in each subgroup, relative to our affecting posterior fossa anatomy were retrieved from our newly developed prenatal nomograms12. digital sonographic archives from the period 2007–2013. The imaging work-up of the studied subgroups was All examinations were performed using a Voluson E8, based on the fetal sonographic assessment of supratentor- Voluson 730 Expert or Voluson 730 Pro (GE Healthcare ial anatomy, cerebellar hemispheres, vermis, brainstem Ultrasound, Milwaukee, WI, USA) ultrasound machine, and posterior fossa fluid spaces, two-dimensional ultra- using a transvaginal probe for vertex presentation and sound measurements of the HB structures, relative to pre- 8–11 a transabdominal one for non-vertex presentation. Cases viously published nomograms , and prenatal magnetic with 3D cranial volumes acquired sagittally were selected resonance imaging (MRI). As a rule, prenatal diagnosis and grouped, according to malformation, into eight cat- was confirmed either by autopsy results or by postnatal egories: cobblestone malformation complex (CMC), for- examination. In two fetuses with RES and in one with merly known as lissencephaly type II, Chiari-II malforma- VD, no postmortem or postnatal studies were performed, tion (C-II), pontocerebellar hypoplasia (PCH), rhomben- but all three had typical prenatal sonographic and MRI cephalosynapsis (RES), Dandy–Walker malformation findings suggestive of their diagnoses. (DWM), vermian dysgenesis (VD), persistent Blake’s pouch cyst (PBPC) and megacisterna magna (MCM). RESULTS CMC was defined by the absence or under-development of cortical sulcation relative to gestational age, associated We identified 43 fetuses diagnosed with CNS malforma- with ventriculomegaly and one or more of the following tions affecting MB and HB anatomy: three with CMC, associated anomalies: ocular dysplasia, encephalocele, sevenwithC-II,twowithPCH,fourwithRES,eightwith corpus callosal dysgenesis and MB-HB anomalies, DWM, seven with VD, six with PBPC and six with MCM. including a small brainstem with a kink in the Mean z-scores of MB and HB measurements are presented mid-pons, small cerebellar hemispheres and vermis, in Table 1. The distribution of the studied parameters in and a dysmorphic tectum. C-II was defined by the affected fetuses is shown relative to the reference graphs12 presence of an open spina bifida associated with collapse in Figures S1–S5. of the posterior fossa cerebrospinal fluid spaces and Application of our new nomograms to the eight downward displacement of HB structures. PCH was subgroups yielded the following results. All fetuses defined by a small-for-gestational-age transcerebellar with CMC showed a prominent elongation of the diameter (TCD) associated with a flattened pontine tectum (TL z-score range, 2.53–6.13) and an extremely belly. RES was defined by a hypoplastic cerebellum with thin pons (APPD z-score range, –6.67 to –4.0) fused hemispheres, absence of vermis and horizontal and (Figures S1 and S3). Ultrasound examination also depicted

Table 1 Midbrain and hindbrain measurement z-scores of 43 fetuses with malformations of the posterior fossa, according to malformation type

Midbrain Hindbrain Malformation n GA at US TL z-score APMD z-score APPD z-score SIVD z-score APVD z-score

CMC 3 22.8 (3.3) 4.00 (1.89) −0.01 (0.53) −5.25 (1.34) −4.81 (3.79) −1.88 (0.69) C-II* 7 25.0 (6.2) 2.93 (1.46) −2.52 (1.32) — — — PCH 2 29.5 (5.4) −2.60 (1.60) −2.45 (0.28) −5.06 (0.74) −5.61 (2.85) −3.02 (0.95) RES 4 25.5 (6.5) 2.96 (1.14) −1.02 (1.76) −3.31 (1.92) N/A N/A DWM 8 20.8 (3.7) 0.42 (0.47) 1.04 (1.61) −1.74 (0.91) −1.51 (2.01) −2.61 (1.60) VD 7 23.4 (3.1) 0.83 (0.59) −0.18 (1.38) −2.79 (0.53) −1.29 (1.32) −1.49 (0.60) PBPC 6 23.7 (3.8) 0.20 (0.35) 0.21 (1.16) −1.44 (0.73) 0.12 (1.06) −0.46 (0.98) MCM 6 28.3 (2.9) 0.24 (0.17) −0.07 (1.03) −0.86 (0.75) −0.58 (0.91) 0.26 (0.39)

Data are given as mean (SD). *Hindbrain biometry could not be measured in fetuses with Chiari-II malformation (C-II) due to a blurred demarcation of the pontine and vermian margins as a result of a collapsed cisterna magna and fourth ventricle. APMD, anteroposterior midbrain diameter; APPD, anteroposterior pons diameter; APVD, anteroposterior vermian diameter; CMC, cobblestone malformation complex; DWM, Dandy–Walker malformation; GA, gestational age; MCM, megacisterna magna; N/A, not applicable; PBPC, persistent Blake’s pouch cyst; PCH, pontocerebellar hypoplasia; RES, rhombencephalosynapsis; SIVD, superoinferior vermian diameter; TL, tectal length; US, ultrasound examination; VD, vermian dysgenesis (Dandy–Walker variant). other MB-HB anomalies in all three fetuses: thick the eight had associated ventriculomegaly and/or callosal echogenic tectal margins, a small dysplastic vermis and a dysgenesis. All eight pregnancies underwent TOP and the kinked brainstem (Figures 1a and b). These pregnancies diagnosis was confirmed by pathological examination. were terminated (TOP) upon parental request and after Three of the seven fetuses with VD (Figure 3b) had a approval of a special ad-hoc committee. The diagnosis TL z-score > 2.0 (Figure S1). In one other fetus the APMD was confirmed by autopsy, being described as compatible was small (Figure S2). An APPD z-score below –2.0 was with Walker–Warburg syndrome in all cases. found in all cases (Figure S3). Six of the fetuses had a low Six of the seven (86%) fetuses with C-II had an z-score (< –2.0) for at least one of the vermian diameters. elongated TL of more than 2SD above the mean for A single 23-week fetus had both SIVD and APVD within gestational age. Five (71%) had a shortened APMD of less the normal range (although both were below the mean for than 2SD below the mean for gestational age (Figures 1c, gestational age), but an upward-rotated vermis lacking S1 and S2). At least one of these abnormal findings (long a primary fissure, a flat fastigium, an enlarged fourth TL and/or short APMD) was found in all seven fetuses ventricle and a normally positioned tentorium. Only one in this group. The diagnosis was confirmed by autopsy fetus demonstrated both a small SIVD and a small APVD in four pregnancies and by prenatal MRI and postnatal associated with a borderline-low TCD. Two fetuses had examination in the three liveborn neonates. a small SIVD, but a normal APVD, while three had a Both cases of PCH showed a reduction of all MB and normal SIVD, but a small APVD. These all had a normal HB measurements (Figures 1d and S1–S5) with APPD TCD. One fetus from this subgroup had associated mild z-scores as low as –4.54 in one and –5.58 in the other. ventriculomegaly. Six underwent TOP (including the case One of the fetuses had associated microcephaly. Both with SIVD and APVD within the normal range) and pregnancies underwent TOP and autopsy confirmed the were described as having a small and dysplastic vermis diagnosis. on postmortem examination. In two, a mild distension In three of the four RES cases there was an elongated of the posterior fossa was mentioned in the pathological TL (z-scores > 2.0) and a small APPD (z-scores < –2.0) report. In the single case which continued to live birth, (Figures S1 and S3). All but one demonstrated severe no postnatal follow-up was available, but prenatal MRI ventriculomegaly (Figure 2). They all underwent TOP. confirmed the ultrasonographic findings. The diagnosis was confirmed by prenatal MRI in two and All six fetuses with PBPC (Figure 3c) had normal by autopsy in the other two. MB-HB measurements (Figures S1–S5 and Table 1). The In two (25%) of the fetuses with DWM (Figure 3a) sonographic diagnosis was confirmed by prenatal MRI in there were high (>2.0) TL and APMD z-scores. In half four of them. All these pregnancies continued to live birth. of the fetuses, the APPD was smaller than 2SD below In five, the cystic structure resolved either during the third the mean, but in all, the APPD was below the mean trimester or by the postnatal ultrasound examination, for gestational age (Figures S1–S3). Either the APVD or showing a normal inferior vermian alignment with the SIVD z-score was < –2.0 in seven of the eight fetuses. posterior brainstem wall. APVD and SIVD z-scores < –2.0 were found in 75% In all six cases of MCM, the MB-HB measurements and 37.5% of cases, respectively (Figures S4 and S5). were within the normal range (Figures S1–S5 and In a single 20-week fetus, both SIVD and APVD were Table 1). The anteroposterior diameters of the cisterna normal; this fetus demonstrated a flat fastigium, caudal magna ranged between 10.5 and 15 mm. The sonographic thinning of the vermis and an elevated tentorium. Four of diagnosis was confirmed by prenatal MRI in three of the

Figure 1 Midbrain and hindbrain malformations as assessed in the mid-sagittal plane using three-dimensional multiplanar reconstruction. (a,b) Two fetuses diagnosed as having cobblestone malformation complex, at 26 weeks (a) and 20 weeks (b), demonstrating similar MB-HB dysmorphism. In both, the tectum (T) is elongated and is orientated nearly horizontally. The anterior and posterior walls of the midbrain (arrowheads) show thick echogenic margins. The hindbrain segment of the brainstem is very thin and demonstrates a typical angulated shape (kinked brainstem) pointed towards the pons (P), indicated by a dashed line at the anterior border of the fourth ventricle (4v). Both fetuses also demonstrate agenesis of the corpus callosum, a dysmorphic and dilated third ventricle (3v) and a small dysplastic vermis (V). (c) Chiari-II malformation at 23 weeks. The tectum is elongated and the anteroposterior midbrain diameter (arrowheads) is thin. The hindbrain components are crowded, showing blurred contours and downward displacement of the cerebellum. (d) Pontocerebellar hypoplasia at 34 weeks. All MB and HD structures are very small. six cases. All the pregnancies continued to live birth. In In cases of CMC, the abnormal MB-HB measurements five fetuses, the distension of the cisterna magna persisted, enabled demonstration of typical features: the enlarged TL but did not show significant enlargement in the following was consistent with an elongated tectum, the low APPD prenatal sonographic examinations. with the kinked dysmorphic brainstem13 (Figure 1a and b) and the small vermian diameters with VD. Our nomograms provided a clear demonstration of the DISCUSSION stretched and flattened MB typical of C-II malformation14, In this study we applied our new nomograms for MB by demonstrating an increased TL and a decreased APMD. and HB dimensions12 to eight different groups of fetuses Of note, HB biometry could not be measured in these with MB-HB malformations. In 41 of the 43 cases, our fetuses due to a blurred demarcation of the pontine and pathological findings were confirmed by prenatal MRI vermian margins as a result of a collapsed cisterna magna or autopsy, and/or by postnatal MRI or examination. and fourth ventricle (Figure 1c). Our new nomograms, which for the first time include Both PCH cases showed very small HB biometric the MB, made it possible to demonstrate mesencephalic parameters, consistent with a hypoplastic pons and abnormalities and to interpret novel features of MB-HB vermis (Figure 1d). Interestingly, we also found a decrease anomalies. in all MB measurements, suggesting that the MB could

Figure 2 Midbrain and hindbrain three-dimensional multiplanar reconstruction (MPR) in a 28-week fetus with a normal cerebellum (a,b) compared with in a 27-week fetus with rhombencephalosynapsis (RES) (c,d). Images on the left (a,c) show coronal and on the right (b,d) mid-sagittal posterior fossa planes. Moving the navigation point (np) in the mid-sagittal plane (b) to the inferoposterior region of the vermis (Ver) results in a coronal image (a) of the vermis, in which it appears as a mid-sagittal, echogenic, ‘ladder-like’ structure, located inside the posterior cerebellar notch, between the hypoechoic cerebellar hemispheres. In RES, the coronal section (c) demonstrates the cerebellar folia running horizontally on the posterior surface of the fused cerebellar hemispheres (Cb) forming continuous echogenic lines. The cerebellum appears as a small oval-shaped structure. P, pons. be involved in PCH (Figures S1–S5). An abnormally is elongated. Indeed, Barkovich and Newton19 described small MB due to CASK mutations has been described a distorted mesencephalic tectum in patients with benign in PCH15. In a neuropathological study of three cases aqueductal narrowing, some of them having an elongated with PCH type 2, Barth et al.16 found a diminished size and thin tectum. The reason for the mesencephalic and cell density of the red nuclei. However, in a large abnormality in our fetuses with RES is unclear, but can study by Namavar et al.17, including 169 cases of PCH, possibly be explained by abnormal MB-HB patterning20. the MB dimensions were not mentioned. We suggest that Of note, parallel shifting of the coronal plane of the MB biometry should be evaluated in future radiological 3D-MPR display (Figure 2c) enabled clear imaging of studies on different types of PCH in order to clarify if the the fused cerebellar hemispheres and an abnormal folial pathological process affecting the HB also involves the pattern in all RES cases. mesencephalon. At least one of the vermian diameters was small in seven We found a small APPD in three of the four of the eight fetuses with DWM. As a result of the upward cases of RES, consistent with involvement of the pons rotation of the vermis, the SIVD was measured nearly (Figure S4) and a high TL in three, implying tectal horizontally and the APVD almost vertically (Figure 3a). elongation (Figure S1). A recent MRI study reported MB Gradual thinning of the caudal vermian edge was observed abnormalities (aqueductal stenosis and midline fusion of in four DWM fetuses, resulting in difﬁcult delineation of the tectum) in about 50% of RES cases and pontine the lower SIVD border. It is interesting that even though hypoplasia in about a quarter of them18. Although the a small vermis is a requirement for the diagnosis of authors did not speciﬁcally address MB length, it is DWM, only two of the eight fetuses demonstrated both possible that, in cases with aqueductal stenosis, the MB short and thin vermis. Four demonstrated only a small

Figure 3 Midbrain and hindbrain sonograms in fetuses with Dandy–Walker continuum as assessed in the mid-sagittal plane using three-dimensional multiplanar reconstruction. (a) Dandy–Walker malformation (DWM) at 20 gestational weeks; (b) vermian dysgenesis (VD) at 27 weeks; (c) persistent Blake’s pouch cyst (PBPC) at 25 weeks. All three cases had normal tectal length (TL) (small arrows) and anteroposterior midbrain diameter (APMD) (arrowheads) measurements. A small dysplastic vermis (V) is indicated in (a) and (b). Due to the upwards vermian rotation in DWM and VD, the superoinferior vermian diameter (SIVD) was measured in a nearly horizontal direction and the anteroposterior vermian diameter (APVD) in an almost vertical one. (c) In PBPC, a small cystic structure balloons through the dilated foramen of Magendie and the vermis shows normal morphology and size. , communication between fourth ventricle and cisterna magna; P, pons.

APVD and one only a small SIVD. It is possible that the a normal SIVD, but a small APVD, thus suggesting an gradual caudal thinning of the vermis caused difficulty in additional form of VD. demarcation of the inferior border, thus giving excessive An interesting finding of our study was that all of the SIVD results. An additional possible explanation is that fetuses with VD also had a small APPD. A flat pons all of our cases were diagnosed in the second trimester; has been well described in PCH but, to the best of our since DWM is a developmental anomaly, the growth of the knowledge, it has not been considered as a part of VD. vermis could be affected progressively and result in smaller The flat pons in PCH is explained by a decreased number measurements by the end of pregnancy. The possibility of pontocerebellar fibers in the developing HB16.This that some vermian measurements in fetuses with DWM explanation could also apply to cases of VD. The small diagnosed in the second trimester may be normal should APPD in VD may be used as an additional feature to be confirmed in a larger study. However, if this is indeed differentiate this from PBPC (Figure 3c). the case, it may imply that, at this stage of gestation, Our new nomograms12 enabled novel characterization if all the other dysmorphic features are present, a small of previously described features of posterior fossa vermis according to SIVD and APVD measurements is not anomalies. We observed an elongated tectum as a feature necessary for diagnosis of DWM. of CMC in addition to a kinked and thin brainstem. A The distinction between vermian hypoplasia and thin and elongated MB reflected downward traction of the agenesis is unclear in both prenatal and postnatal brainstem in C-II. The MB showed reduced dimensions literature21. In fact, the terms ‘hypoplasia’ and ‘agenesis’ in PCH. The elongated tectum in RES can represent are used interchangeably in the literature on fetal the MB anomaly leading to aqueductal stenosis and vermian abnormalities22. Furthermore, the term ‘vermian hydrocephalus. The pons was involved in VD (albeit dysgenesis’ or ‘dysplasia’ has been used when the lobules less so than in PCH). Furthermore, the nomograms of the vermis are abnormally formed, regardless of enabled differentiation between the various forms of their number or size23. Our results also confirmed the VD, either in isolation or as part of DWM; and normal difficulty in prenatal differentiation between hypoplasia, vermian and pontine biometry distinguished VD from agenesis (which is usually partial) and dysgenesis of the PBPC and MCM. vermis. All of the VD fetuses had a slightly upwardly These findings should be confirmed in future prenatal rotated dysmorphic vermis with a flat fastigium and and postnatal studies in larger groups of patients with undetectable primary fissure. This could not be defined posterior fossa malformations. as hypoplastic, since vermian hypoplasia is defined as a small but normally formed vermis. It also could not be defined as partial agenesis, since this requires that part of REFERENCES the vermis is missing but the rest is normally formed21. 1. Barkovich AJ. Developmental disorders of the midbrain and We therefore defined it as VD. It may be assumed that hindbrain. Front Neuroanat 2012; 6: 1–10. in cases with vermian hypoplasia, both measurements 2. Chizhikov V, Millen KJ. Development and malformations of (SIVD and APVD) will be low, while in partial agenesis the cerebellum in mice. Mol Genet Metab 2003; 80: 54–65. the APVD should be normal, but the SIVD should be 3. Senturk¨ A, Pfennig S, Weiss A, Burk K, Acker-Palmer A. Ephrin small due to a truncated shape of the inferior vermis. Two Bs are essential components of the Reelin pathway to regulate neuronal migration. Nature 2011; 472: 356–360. of our fetuses depicted a small SIVD, but a normal APVD, 4. Li S, Jin Z, Koirala S, Bu L, Xu L, Hynes RO, Walsh CA, Corfas implying partial agenesis; one had both a small SIVD and G, Piao X. GPR56 regulates pial basement membrane integrity a small APVD, suggesting hypoplasia; three demonstrated and cortical lamination. J Neurosci 2008; 28: 5817–5826.

5. Doherty D. Joubert syndrome: insights into brain development, Zanni G, Petit F, Afenjar A, Goizet C, Barresi S, Coussement A, cilium biology, and complex disease. Semin Pediatr Neurol Ioos C, Lazaro L, Joriot S, Desguerre I, Lacombe D, des Portes 2009; 16: 143–154. V, Bertini E, Siffroi JP, de Villemeur TB, Rodriguez D. Spectrum 6. Soto-Ares G, Yjoyes B, Lemaitre M, Vallee L, Pruvo J. MRI of pontocerebellar hypoplasia in 13 girls and boys with CASK in children with mental retardation. Pediatr Radiol 2003; 33: mutations: confirmation of a recognizable phenotype and first 334–345. description of a male mosaic patient. Orphanet J Rare Dis 7. Barkovich AJ, Millen KJ, Dobyns WB. A developmental and 2012; 7: 18. genetic classification for midbrain-hindbrain malformations. 16. Barth PG, Aronica E, de Vries L, Nikkels PG, Scheper W, Hooze- Brain 2009; 132: 3199–3230. mans JJ, Poll-The BT, Troost D. Pontocerebellar hypoplasia 8. Achiron R, Kivilevitch Z, Lipitz S, Gamzu R, Almog B, Zalel Y. type 2: a neuropathological update. Acta Neuropathol 2007; Development of the human fetal pons: in utero ultrasonographic 114: 373–386. study. Ultrasound Obstet Gynecol 2004; 24: 506–510. 17. Namavar Y, Barth PG, Kasher PR, van Ruissen F, Brockmann 9. Malinger G, Ginath S, Lerman-Sagie T, Watemberg N, Lev D, K, Bernert G, Writzl K, Ventura K, Cheng EY, Ferriero DM, Glezerman M. The fetal cerebellar vermis: normal development Basel-Vanagaite L, Eggens VR, Krageloh-Mann¨ I, De Meirleir as shown by transvaginal ultrasound. Prenat Diagn 2001; 21: L, King M, Graham JM Jr, von Moers A, Knoers N, Sztriha 687–692. L, Korinthenberg R; PCH Consortium, Dobyns WB, Baas F, 10. Ginath S, Lerman-Sagie T, Haratz Krajden K, Lev D, Poll-The BT. Clinical, neuroradiological and genetic findings in Cohen-Sacher B, Bar J, Malinger G. The fetal vermis, pons pontocerebellar hypoplasia. Brain 2011; 134: 143–156. and brainstem: normal longitudinal development as shown 18. Ishak GE, Dempsey JC, Shaw DW, Tully H, Adam MP, by dedicated neurosonography. J Matern Fetal Neonatal Med Sanchez-Lara PA, Glass I, Rue TC, Millen KJ, Dobyns WB, 2013; 26: 757–762. Doherty D. Rhombencephalosynapsis: a hindbrain malforma- 11. Mirlesse V, Courtiol C, Althuser M; CFEF (College` français tion associated with incomplete separation of midbrain and d’echographie´ foetale), Duyme M. Ultrasonography of the forebrain, hydrocephalus and a broad spectrum of severity. fetal brainstem: a biometric and anatomical, multioperator, Brain 2012; 135: 370–386. cross-sectional study of 913 fetuses of 21–36 weeks of gestation. 19. Barkovich AJ, Newton TH. MR of aqueductal stenosis: Prenat Diagn 2010; 30: 739–745. evidence of a broad spectrum of tectal distortion. AJNR Am J 12. Leibovitz Z, Shkolnik C, Krajden Haratz K, Malinger G, Neuroradiol 1989; 10: 471–476. Shapiro I, Lerman-Sagie T. Assessment of fetal midbrain and 20. Yachnis AT. Rhombencephalosynapsis with massive hydro- hindbrain in mid-sagittal cranial plane by three-dimensional cephalus: case report and pathogenetic considerations. Acta multiplanar sonography. Part 1: comparison of new and Neuropathol 2002; 103: 301–304. established nomograms. Ultrasound Obstet Gynecol 2014; 44: 21. Lerman-Sagie T, Malinger G, Ben-Sira L. Prenatal diagnosis of 575–580. structural brain anomalies. In Swaiman’s Pediatric Neurology th 13. Stroustrup Smith A, Levine D, Barnes PD, Robertson RL. (5 edn), Swaiman K, Ashwal S, Ferriero DM, Schor NF (eds). Magnetic resonance imaging of the kinked fetal brain stem: Elsevier Inc: Edinburgh, 2012; 263–276. a sign of severe dysgenesis. J Ultrasound Med 2005; 24: 22. Guibaud L, des Portes V. Plea for an anatomical approach 1697–1709. to abnormalities of the posterior fossa in prenatal diagnosis. 14. Shoja MM, Tubbs RS, Oakes WJ. Embryology and pathophys- Ultrasound Obstet Gynecol 2006; 27: 477–481. iology of the Chiari I and II malformations. In The Chiari 23. Russo R, Fallet-Bianco C. Isolated posterior cerebellar vermal Malformations, Tubbs RS, Oakes WJ (eds). Springer: New defect: a morphological study of midsagittal cerebellar vermis York, US, 2013; 55–73. in 4 fetuses--early stage of Dandy-Walker continuum or new 15. Burglen L, Chantot-Bastaraud S, Garel C, Milh M, Touraine R, vermal dysgenesis? J Child Neurol 2007; 22: 492–500.

SUPPORTING INFORMATION ON THE INTERNET

The following supporting information may be found in the online version of this article: Figures S1–S5 Individual values of tectal length (TL) (Figure S1), anteroposterior midbrains diameter (APMD) (Figure S2), anteroposterior pons diameter (APPD) (Figure S3), superoinferior vermian diameter (SIVD) (Figure S4) and anteroposterior vermian diameter (APVD) (Figure S5), according to gestational age in fetuses diagnosed with central nervous system malformations affecting posterior fossa development (cobblestone malformation complex, Chiari-II malformation, pontocerebellar hypoplasia, rhombencephalosynapsis, Dandy–Walker malformation, vermian dysgenesis, persistent Blake’s pouch cyst and megacisterna magna), plotted on normal reference ranges which were modeled according to second-order polynomial trends12: 2 2 tectal length (TL mean+2SD = –0.0007 × GW + 0.0606 × GW + 0.0123; TL mean = –0.0007 × GW + 0.0556 2 × GW – 0.0396; TL mean–2SD = –0.0007 × GW + 0.0507 × GW – 0.0916); anteroposterior midbrain 2 2 diameter (APMDmean+2SD = 0.0017 × GW – 0.0346 × GW + 1.2799; APMDmean = 0.0006 × GW + 0.0113 2 × GW + 0.6107; APMDmean-2SD = –0.0005 × GW + 0.0571 × GW – 0.0584); anteroposterior pons 2 2 diameter (APPDmean+2SD = 0.0003 × GW + 0.0536 × GW – 0.2223; APPDmean = 0.0003 × GW + 0.0406 2 × GW – 0.1135; APPDmean-2SD = 0.0002 × GW + 0.0277 × GW – 0.0046); superoinferior vermian diameter 2 2 (SIVDmean+2SD = –0.0002 × GW + 0.0981 × GW – 0.6965; SIVDmean = –0.0009 × GW + 0.1226 2 × GW – 1.0454; SIVDmean-2SD = –0.0017 × GW + 0.147 × GW – 1.3942); anteroposterior vermian 2 2 diameter (APVDmean+2SD = –0.0007 × GW + 0.1054 × GW – 0.9674; APVDmean = –0.0008 × GW 2 + 0.0932 × GW – 0.8607; APVDmean+2SD = –0.001 × GW + 0.081 × GW – 0.7539). All measurements are expressed in cm.