The Center for Cerebellar Malformations
Dandy Walker Malformation
General Information
The
Dandy-Walker malformation was first described in 1887 by Sutton (Sutton, 1887) and was further characterized by Dandy and Blackfan
in 1914 and Taggart and Walker in 1942 (Dandy and Blackfan, 1914; Taggart and Walker, 1942). The key components of this malformation include
hypoplasia of the cerebellar vermis and cystic dilatation of the 4th ventricle.
The 4th ventricle communicates with a retrocerebellar cyst that may cause
enlargement of the posterior fossa and elevation of the tentorium, seen on
imaging studies as elevation of the torcula or confluence of the sinuses. A
third, variable component of DWM is communicating hydrocephalus with enlarged
lateral ventricles.
Presentation
of DWM can take many forms. It
often presents with macrocephaly in the neonatal period, and infants may come
to medical attention because of hydrocephalus, developmental delay, or ataxia. It may be identified on prenatal
ultrasound, where prenatal counseling takes a critical role. It may be present in asymptomatic
individuals undergoing brain imaging for an unrelated reason.
DWM is
a relatively common malformation, occurring in at least 1 in 5000 liveborn
infants (Parisi and Dobyns, 2003). It has
been reported in a wide variety of chromosomal anomalies, and in several
different genetic syndromes (Bordarier and Aicardi, 1990; Chitayat et al., 1994;
Murray et al., 1985). However, some of these
syndromes, specifically Meckel-Gruber and Walker-Warburg syndromes, have
complex mid-hindbrain malformations that are unlikely to represent the classic
Dandy-Walker malformation. Some surveys suggest that environmental factors,
including prenatal exposure to teratogens such as rubella or alcohol, are
associated with DWM (Benke, 1984; Clarren et al., 1978).
Diagnostic criteria and associated features
The
core criteria of Dandy-Walker malformation are: 1. cerebellar vermis
hypoplasia; 2. cystic enlargement of the 4th ventricle; and3. elevation of the
roof of the posterior fossa (the tentorium cerebelli and torcula (Parisi and Dobyns, 2003) (see MRI below). In addition to these diagnostic criteria, there is often
enlargement of the posterior fossa, stenosis of the outflow tracts of the 4th
ventricle, and hydrocephalus with increased intracranial pressure. Associated
malformations, generally central nervous system in origin (including occipital
encephalocele, polymicrogyria, and heterotopia), are present in 25-50% of
individuals with Dandy-Walker malformation. A significant proportion (10-17%)
display agenesis or dysgenesis of the corpus callosum (Ben Hamouda et al., 2001) (Wilson et al., 1994). Other
non-CNS anomalies with an increased frequency in DWM include congenital heart
disease, cleft lip and/or palate, and neural tube defects (Huong et al., 1975; Ritscher et al., 1987). A
recurring association of DWM with facial hemangiomas has been noted and
described under the acronym PHACE syndrome (Posterior fossa brain
malformations, Hemangiomas, Arterial anomalies, Coarctation of the aorta and
cardiac defects, and Eye abnormalities) (Frieden et al., 1996; Poetke et al., 2002).
MRI Scans
Dandy-Walker malformation.
Dandy-Walker Malformation. (A) Note the presence of the cerebellar
vermis (asterisk) that is rotated upwards. A cystic dilatation of the fourth ventricle replaces the
site normally taken by the vermis.
(B) Axial image through the pons.
The vermis is visible (white arrow), and there is no molar tooth
malformation (black arrows).
Conditions
to be distinguished from Dandy-Walker.
Conditions similar to Dandy-Walker. A: Cerebellar vermis
hypoplasia/atrophy. Note normal
size of posterior fossa and absence of hydrocephalus. B. Retrocerebellar cyst with shift of cerebellum
anteriorly. This results in
obstruction of CSF outflow and resultant hydrocephalus. C: Mega cisterna
magna. Note enlarged size of
posterior fossa, but normal size of cerebellum. D. BlakeÕs pouch cyst.
Note CSF collection in IVth ventricle that is contiguous with a
collection inferior to the cerebellum (*). Small arrows show upward mass effect from fluid. In none of these conditions is the
fourth ventricle significantly enlarged or upwardly rotated.
Variability
of the Dandy-Walker Malformation
There are at least four cerebellar malformations that are frequently confused with DWM: cerebellar vermis hypoplasia/aplasia, retrocerebellar cyst, mega cisterna magna, and BlakesÕ pouch cyst. Collectively, these have been grouped together as Dandy-Walker variants, but there is a growing realization that they have no direct relation to DWM and should be treated as distinct from DWM. The confusion initially arose when axial-plane CT was the standard imaging modality (Archer et al., 1978), because these different malformations are best appreciated in the parasagittal plane. Cerebellar vermis hypoplasia/atrophy can be observed as an isolated finding, or may be part of a metabolic or genetic defect, such as associated with congenital disorders of glycosylation (CDG) (Worthington et al., 1997) or with mutations in the oligophrenin gene (des Portes et al., 2004). Retrocebellar cysts, mega cisterna magna and BlakeÕs pouch cysts may be part of a continuum, and cannot always be distinguished from one another, but each can usually be distinguished from DWM. A BlakeÕs pouch cyst is merely an extension of the 4th ventricle through the foramen of Magendie into the vallecula. It may be associated with mild vermian hypoplasia and may be a forme fruste of the Dandy-Walker malformation. Retrocerebellar cysts are usually arachnoid cysts that sit in the fairly dense subarachnoid space posterior to the vermis. They are usually more rostral and more dorsal than a BlakeÕs pouch cyst. Mega cisterna magna is applied to large retrocerebellar CSF collections in patients who are asymptomatic or are symptomatic from other conditions. Often, these posterior fossa fluid collections are of no clinical significance (Boltshauser et al., 2002), unless associated with cerebellar parenchyma volume loss, CSF fluid obstruction resulting in hydrocephalus, or cerebellar compression.
Clinical course
DWM
often present in the neonatal period with macrocephaly, occipital cephaloceles,
and/or hydrocephalus (Hirsch et al., 1984). For
those with severe obstructive hydrocephalus, multiple congenital anomalies,
and/or other severe CNS anomalies such as porencephaly, the mortality is high.
Apnea and seizures are seen in a significant proportion of children with DWM (Gerszten and Albright, 1995), although developmental delay and mental retardation
are highly variable. These infants tend to have congenital hypotonia and may
later develop spasticity (Maria et al., 1987). Ataxia
and nystagmus are seen in many, but cerebellar signs are variable and may not
be present. Many subjects are diagnosed in infancy, due to increasing head
circumference and/or symptoms of elevated intracranial pressure such as
lethargy, vomiting, and irritability; however, some will display normal
intellect and essentially normal motor function(Maria et al., 1987). There are reports of DWM diagnosed incidentally
after cranial imaging studies performed for other indications (Lipton et al., 1978).
Treatment
The
treatment of DWM has been a subject of great controversy. In early series,
based on the belief that hydrocephalus was due to obstruction of the foramina
of Lushka and Magendie, surgery involved excision of the posterior fossa membranes
to create unobstructed flow of CSF, with resultant poor outcomes (Sawaya and McLaurin, 1981). Treatment by either direct shunting of the lateral ventricles
in the presence of hydrocephalus is typically required but has met with mixed
success (Kumar et al., 2001). Although it has been proposed that return of normal
cerebellar architecture by shunting the cyst is associated with good functional
outcome (Golden et al., 1987), other authors suggest that the measured volume of
cerebellum is not significantly changed by cyst shunting and advocate
ventriculoperitoneal shunting as the best approach to relieve increased
intracranial pressure (Gerszten and Albright, 1995). Cognitive outcomes in DWM series vary widely and
appear to fit a bimodal distribution, suggesting that factors other than the
malformation itself contribute to the outcome (Parisi and Dobyns, 2003).