Posterior fossa anomalies diagnosed with fetal MRI: Associated anomalies and
A thesis submitted to the
Graduate School of the University of Cincinnati
in partial fulfillment of the requirements
for the degree of
Master of Science
In the Program of Genetic Counseling
of the College of Medicine
Kyla Patek, B.S.
University of Minnesota, 2009
May 9th 2011
Committee Chair: Robert J. Hopkin, M.D.
Objective: The purpose of this study was to describe the relationship between intra- and extra-
cranial anomalies and neurodevelopmental outcome for fetuses diagnosed with a posterior fossa
anomaly (PFA) on fetal MRI.
Methods: Cases of Dandy-Walker malformation (DWM), vermian hypogenesis/hypoplasia
(VH), and mega cisterna magna (MCM) were identified through the Fetal Care Center of
Cincinnati between January 2004 and December 2010. Parental interview and retrospective
chart review were used to assess neurodevelopmental outcome.
Results: PFAs were identified in 59 fetuses; 9 with DWM, 36 with VH, and 14 with MCM.
Cases with isolated PFAs (14/59) had better outcomes than those with additional anomalies
(p=0.00016), with isolated cases of MCM all being neurodevelopmentally typical. Cases with
additional intra-cranial anomalies were more likely to have a poor outcome than those without
intra-cranial anomalies (p=0.00085). The presence of extra-cranial anomalies increased the
likelihood of having a poor outcome (OR=10.2, 95% CI, 2.6-40) as did the identification of an
abnormal brainstem (OR=12.97, 95% CI, 2.9-58.0).
Conclusion: Intra- and extra-cranial anomalies were good predictors of neurodevelopmental
outcome in this study. The prognosis was poor for individuals with an abnormal brainstem while
those with isolated MCM had normal neurodevelopmental outcome.
Key Words: Posterior fossa anomaly; fetal MRI; neurodevelopmental outcome, Dandy-Walker
malformation, vermian hypoplasia, mega cisterna magna
Thank you to my RAC (Christine Spaeth, Dr. Robert Hopkin, and Beth Kline-Fath) for
their support, collaboration, and guidance.
Table of Contents
List of Figures and Tables vi
List of Figures and Tables
Figure 1. Fetal MRI of a Dandy-Walker malformation with an abnormal brainstem at 32 weeks
Figure 2. Fetal MRI of vermian hypoplasia at 33 weeks gestation.
Figure 3. Fetal MRI of mega cisterna magna at 31 weeks.
Figure 4. Fetal MRI of vermian hypoplasia with a small brainstem at 24 weeks gestation.
Figure 5. Neurodevelopmental outcome for cases with vermian hypogenesis/hypoplasia (VH).
Table 1. Outcome of posterior fossa anomalies.
Table 2. Neurodevelopmental outcome for 6 surviving cases with Dandy-Walker malformation.
Table 3. Neurodevelopmental outcome for 21 surviving cases with vermian
Table 4. Neurodevelopmental outcome for 10 cases with mega cisterna magna.
Posterior fossa anomalies (PFAs), including Dandy-Walker malformation (DWM),
vermian hypogenesis/hypoplasia (VH), and mega cisterna magna (MCM), are found in
approximately 1 in every 5,000 live births (Bolduc and Limperopoulos, 2009). The
identification of a PFA leads to a thorough ultrasound exam or fetal MRI when available. Fetal
MRI provides detailed imaging of the soft tissue of the brain to help differentiate between DWM,
VH, and MCM, which have different prognoses (Limperopoulos and du Plessis, 2006,
Adamsbaum et al., 2005). Fetuses suspected of having a PFA are at an increased risk for both
additional intra- and extra-cranial anomalies. Several studies have reported the importance of the
identification of additional anomalies for providing accurate prognoses to families (Cornford and
Twining, 1992, Estroff et al., 1992, Ecker et al., 2000).
Neurologic complications and developmental delay are commonly seen in children with
PFAs. Many studies have reported the rate of developmental delay in children identified with a
PFA ranging from 20-80% depending on the specific anomaly and case series (Klein et al., 2003,
Bolduc and Limperopoulos, 2009, Gerszten and Albright, 1995, Ecker et al., 2000,
Limperopoulos and du Plessis, 2006). The lack of evidence-based outcome data for fetuses
diagnosed with PFAs makes it difficult to provide prognostic information to families receiving a
The purpose of this study was to describe the relationship between intra- and extra-
cranial anomalies and neurodevelopmental outcome for fetuses diagnosed with a posterior fossa
anomaly (PFA) on fetal MRI consecutively between January 2004 and December 2010 from a
single quaternary care center.
After receiving Institutional Review Board approval, a retrospective chart review was
conducted at the Fetal Care Center of Cincinnati (FCC) at Cincinnati Children’s Medical Center.
All fetal MRIs were reviewed by a radiologist to assess accuracy of diagnosis and identify
additional defects. Postnatal images were reviewed when available. Diagnosis of
Dandy-Walker malformation was made in the presence of vermian agenesis/hypogenesis, cystic
dilatation of the fourth ventricle and enlarged posterior fossa with elevation of the transverse
sinus, tentorium and torcula (Figure 1). Vermian abnormalities were described based on
standardized measurements described by Garel (2009) and Robinson et al (2001). Vermian
hypogenesis was noted if the vermis measured small and the fastigial to declive line was less
than a 1:2 ratio. The diagnosis of vermian hypoplasia was verified in the presence of a small
vermis with normal foliation (Figure 2). Due to the sample size, vermian hypogenesis and
hypoplasia were treated as one group for the purpose of analysis. The brainstem was identified
as abnormal if small or irregular in configuration compared to normal brainstem measurement
standards for gestational age described by Tilea et al (2009) (Figure 4).
Data was collected from medical records (prenatal and postnatal) (Appendix 1) with
additional information collected through a voluntary parental phone interview (Appendix 2).
Women who were referred to FCC for evaluation but experienced a prenatal or perinatal demise
were not contacted for the phone interview but were included in the chart review. Data collected
from medical record review and phone surveys included the outcome of the pregnancy (delivery,
stillbirth, intrauterine fetal demise or termination), medical history of the pregnancy, postnatal
imaging, genetic diagnosis/testing, neurologic abnormalities (i.e. seizures, hyper/hypotonia,
respiratory distress, feeding difficulties), developmental milestones and parental assessment of
development. Developmental milestones were obtained from parental report and medical
records. Developmental function was estimated as mild, moderate or severe based on guidelines
provided in the Denver Developmental Screening Test II or other validated developmental
testing recorded in the medical records. For statistical purposes outcome was classified as good
if the individual had no neurologic abnormalities or development delay and poor if the
individuals had neurologic abnormalities, developmental delay, or died.
Fisher Exact test was employed to examine the association between specific features
found on MRI and postnatal outcome. Because there were 4 explanatory variables (isolated
PFA, additional intra-cranial anomalies, extra-cranial anomalies, and abnormal brainstem) and
multiple comparison tests were performed, Bonferroni correction with ? ≤ 0.005 was used.
Fisher Exact test, where ? ≤ 0.05, was used to assess the accuracy of fetal MRI based on
gestational age. A proportional odds model was developed to determine the strength of
association between fetal brain malformations and outcome. Inclusion of all explanatory
variables in the model resulted in statistically insignificant coefficients because of overlap
between groups. Odds ratios were estimated for extra-cranial anomalies and abnormal
A total of 59 fetuses were diagnosed with a PFA on fetal MRI (Table 1); 9 (15%) had
DWM, 36 (61%) had VH and 14 (24%) had MCM. The average gestational age at the time of
fetal MRI was 25.6 weeks (±4.8 weeks). Postnatal brain imaging was available for confirmation
in 29 cases, and the original diagnosis was confirmed in 24 of the 29 cases. All 5 unconfirmed
cases had VH diagnosed before 24 weeks gestation. For VH, diagnoses made after 24 weeks
were significantly more likely to be confirmed compared to those made before 24 weeks
(p=.033). A genetic abnormality was identified by chromosome analysis or high density
microarray in 9 cases (3 with DWM, 5 with VH, and 1 with MCM). All individuals with a
genetic abnormality were non-isolated cases.
Of the 9 cases with DWM, 2 were electively terminated and 1 died neonatally. All 6
survivors had developmental and neurologic abnormalities (Table 2). The mean age at the time
of the study was 2.3 years old (range: 2 months to 4.5 years). DWM was isolated in 1 case,
occurred with only additional extra-cranial anomalies in 1 case, and occurred with both intra-
and extra-cranial anomalies in 7 cases.
Vermian abnormalities were identified in 36 cases. Vermian hypogenesis was seen in 4
individuals (cases 5, 15, 18 and 27), while the remaining 32 cases were diagnosed with vermian
hypoplasia. Neonatal death occurred in 10 cases, 1 was electively terminated, and 2 were lost to
follow up. Additionally, 2 individuals died during the first year of life, leaving 21 cases with
postnatal information available (Table 3). Of the 21 living cases, neurodevelopmental outcome
varied greatly (Figure 5). Five cases had both additional intra-and extra-cranial anomalies; all
had abnormal neurodevelopmental outcome.
Mega cisterna magna was identified in 14 cases. Neurodevelopmental outcome
information was available for 10 cases (Table 4) and 4 cases were lost to follow up. The MCM
was isolated in 7 cases, all of which had normal neurologic and developmental outcomes. Extra-
cranial anomalies were identified in 3 cases with MCM; all 3 had neurologic abnormalities and 2
out of 3 also had developmental delay.
The strength of association between PFAs and neurodevelopmental outcome was
examined for 50 of the 59 cases under the following conditions: isolated or when seen in the
presence of intra-cranial abnormalities, extra-cranial abnormalities, or an abnormal brainstem.
Those lost to follow up (6/59) or terminated (3/59) were excluded. The average age at the time
neurodevelopmental outcome was gathered was 2.9 years old (range: 1 month to 6.5 years).
Cases with isolated PFAs had statistically better neurodevelopmental outcomes than cases
identified with additional anomalies (14/50 and 36/50, respectively; p=0.00016). Cases with
additional intra-cranial anomalies were more likely to have a poor outcome compared to those
without additional intra-cranial anomalies (26/50 and 24/50 respectively; p=0.00085). Similarly,
the presence of extra-cranial anomalies was associated with a worse outcome compared to those
without extra-cranial anomalies (30/50 and 20/50 respectively; p=0.00014). The presence of
extra-cranial anomalies with a PFA increased the likelihood of having a poor
neurodevelopmental outcome or death (OR=10.2, 95% CI, 2.6-40). Brainstem abnormalities
were identified in 15 cases (30%); of these, 9 died and 6 have neurologic abnormalities and
severe developmental delay. The presence of a brainstem abnormality was related to a
significantly worse outcome compared to those with normal brainstems (15/50 and 35/50
respectively; p=0.00018) and increased the odds of having a poor neurodevelopmental outcome
or death (OR=12.97, 95% CI, 2.9-58.0).
Posterior fossa anomalies are a relatively common finding on prenatal ultrasound. The
variation in neurodevelopmental outcome seen between different PFAs as well as within each
specific anomaly is great, making counseling families receiving this diagnosis challenging.
Many studies have reported a relationship between additional intra- and extra-cranial anomalies
and neurodevelopmental outcome (Kolble et al., 2000, Ecker et al., 2000, Forzano et al., 2007,
Bolduc and Limperopoulos, 2009). In this study, the identification of additional intra-cranial
anomalies or extra-cranial anomalies was associated with a worse outcome regardless of the type
of PFA. These results support previous studies in emphasizing the importance of a full anatomic
scan to establish whether additional intra- or extra-cranial anomalies are present (Levine et al.,
1999, Adamsbaum et al., 2005). Additionally, fetal MRI provides detailed information about
neuroanatomy and can identify brain anomalies, such as heterotopia or abnormal gyria that could
be missed on a level II ultrasound. Considering 20% of cases of DWM, VH, and non-isolated
MCM had a chromosome abnormality detected, an amniocentesis with chromosome analysis and
high density microarray should be considered in the diagnostic work-up for a fetus identified
with a PFA.
Of children who had a poor outcome, the odds that they had a brainstem abnormality
were greater than the odds that they had an extra-cranial abnormality. To the knowledge of the
authors, a statistical analysis of the relationship between brainstem abnormalities and outcome
for individuals with a PFA has not been previously reported in the literature. Although previous
studies have not treated brainstem abnormalities as a separate risk factor from other anomalies,
our analysis demonstrates that when abnormal, the brainstem is important for predicting
neurodevelopmental outcome. All brainstem abnormalities in this study led to a poor outcome,
with a significant number of neonatal deaths. The literature on specific syndromes involving
brainstem abnormalities, such as Walker Warburg syndrome, indicates that the outcome is poor
when the brainstem is involved (Vajsar and Schachter, 2006). It therefore is important that
families should be given a guarded prognosis when a brainstem abnormality is found in addition
to a PFA, regardless of whether additional anomalies or a specific syndrome is identified.
In previous literature, neurodevelopmental delay occurs in 40-60% of children with
DWM, with the identification of additional anomalies leading to a worse prognosis (Klein et al.,
2003, Boddaert et al., 2003). In this study all but one case had additional intra- and/or extra-
cranial anomalies and all living children had hydrocephalus which required a shunt. These
findings suggest that a DWM is more likely to be found in the presence of additional anomalies
rather than as an isolated occurrence. Similar to previous studies where high rates of additional
anomalies were found (Long et al., 2006), all cases of DWM in this study had a poor outcome.
Unlike DWM, classification of vermian abnormalities has changed rapidly in the last ten
years leading to difficulties in making meaningful comparison between studies looking at
vermian hypoplasia, Dandy-Walker variant, Dandy-Walker continuum, and Dandy-Walker
complex. Although there were not enough cases in our study population to perform statistical
analysis, there did not appear to be a difference in outcome between cases with vermian
hypogenesis versus vermian hypoplasia. Postnatal confirmation of VH was made in 83% of
cases, demonstrating that false positive diagnoses can still occur when the fetal MRI is
performed before 24 weeks. This is similar to previously reported confirmation rates, and is
thought to be due to normal variation in development (Limperopoulos et al., 2006). Many
studies have reported a better prognosis for individuals with isolated VH compared to individuals
with VH and additional anomalies (Bolduc and Limperopoulos, 2009, Harper, 2007); however,
8 Download full-text
isolated VH did not have a better prognosis in this study. Despite the insignificance of VH with
intra- or extra-cranial anomalies separately, when both were present the outcome was universally
poor, with all having either poor neurodevelopmental outcome or death. These findings are
similar to a previous report that risk for death increases significantly when additional anomalies
are found in two or more organ systems (Salihu et al., 2008).
The question of whether an isolated MCM (≥10 mm) is pathogenic or a benign variant in
human development has important implications for parents who are be considering termination.
This study adds to the current literature which is consistent with a normal neurodevelopmental
postnatal outcome for fetuses with an isolated MCM (Bolduc and Limperopoulos, 2009,
Adamsbaum et al., 2005, Long et al., 2006, Haimovici et al., 1997, Dror et al., 2009, Forzano et
al., 2007). Previous reports indicate that non-isolated cases of MCM have good outcomes in 11-
29% of cases (Forzano et al., 2007, Long et al., 2006). The non-isolated cases of MCM in this
study had a worse outcome than has been reported in the literature, with all having neurologic
abnormalities and/or developmental delay. However, this group only included three cases and
therefore may not represent the true variation in neurodevelopmental outcome that exists in the
An important limitation of this study was the lack of standardized evaluation of
neurodevelopmental outcome. In a review of studies looking at neurodevelopmental outcome in
children with cerebellar malformations, 74% of studies shared this same limitation (Bolduc and
Limperopoulos, 2009). Although both parental report and medical records review were utilized
to assess neurodevelopmental outcome, it is possible that subtle deficits that have been reported
in adult studies could have been missed (Zimmer et al., 2007).