Neurodegenerative Diseases and Hydrocephalus

Colloid Cyst

The classic cause of obstruction at the foramina of Monro is the colloid cyst ( Fig. 7-2 ). This is typically located in the anterior region of the third ventricle. On unenhanced CT, the lesion is high in density. Magnetic resonance (MR) often shows a lesion that is high intensity on T1-weighted imaging (T1WI) and T2WI. The signal of colloid cysts is variable, depending on the protein concentration, presence of hemorrhage, and other paramagnetic ion effects.

Congenital Aqueductal Stenosis

The cerebral aqueduct is one of the narrowest channels through which the CSF in the ventricles must flow. Congenital aqueductal stenosis is just one of the obstructers of the aqueduct. This is most commonly an X-linked recessive disorder seen in early childhood, although it can present at any age. Children typically have enlarging head circumferences and dilatation of the lateral and third ventricles, but with a normal-appearing fourth ventricle. Aqueductal webs, septa, or diaphragms may also obstruct the exit of CSF from the third ventricle. Brain stem, tectal, and pineal region lesions can also result in aqueductal stenosis caused by extrinsic mass effect on the aqueduct.

Sagittal MR is very helpful for distinguishing extrinsic mass compression from an intrinsic aqueductal abnormality ( Fig. 7-3 ). Aqueductal stenosis may also be diagnosed on CSF flow (phase contrast) MR imaging. Phase contrast MR with a velocity encoding set to 10 to 15 cm/sec may be the best way to assess aqueductal patency. Application of two gradients of equal magnitude but opposite direction can produce signal from moving protons moving in either direction, whereas stationary protons do not produce signal. Biphasic flow indicated by sequentially bright and dark signal (indicating to and fro flow) should be seen on the CSF flow scan ( Fig. 7-4 ).

Clots or Synechiae

Other intrinsic causes of ventricular obstruction include clots or synechiae resulting from trauma or chronic infection. Patients with a large amount of subarachnoid hemorrhage may demonstrate obstruction caused by clot formation anywhere within the ventricular system. Synechiae may be due to fibrous adhesions after ventriculitis or meningitis. An infectious cause of ventricular obstruction is cysticercosis. Occasionally, only parts of the ventricular system may be affected, resulting in compartmentalized dilatation of the obstructed components of the ventricular system.

Masses or Tumors

Masses of the pineal gland are the most common causes of extrinsic obstruction of the aqueduct. These generally compress the aqueduct from posteriorly and cause dilatation of the lateral and third ventricular system. Tectal gliomas also obstruct the aqueduct early in their course ( Fig. 7-5 ). Occult cerebrovascular malformations may occur there as well. The base of the aqueduct may be obstructed by tumors of the posterior fossa such as medulloblastomas (primitive neuroectodermal tumors [PNETs]) or ependymomas.

Any of the pediatric and adult posterior fossa tumors may obstruct the fourth ventricle and/or lower portion of the aqueduct. Ependymomas are one of the classic intraventricular tumors to infiltrate the foramina of Magendie and Luschka and may cause hydrocephalus from outflow obstruction. Cerebellar astrocytomas, medulloblastomas, or hemangioblastomas may compress the fourth ventricle extrinsically.

Other tumors that may obstruct parts of the ventricular system include meningiomas, neurocytomas, astrocytomas, choroid plexus papillomas, oligodendrogliomas, subependymal giant cell tumors, arachnoid/ependymal cysts, craniopharyngiomas, and (epi)dermoids.

Hematomas and Infarcts

A well-placed hematoma can compress the ventricles and lead to occlusion at the foramen of Monro, aqueduct, or fourth ventricle. These can be seen in the settings of trauma, acute subdural and/or epidural posterior fossa hematomas, and with hypertensive bleeds. Posterior fossa strokes are notorious for bringing about the downfall of the patient by eliciting acute hydrocephalus as the fourth ventricle is compressed and obliterated by mass effect. At the same time, excessive mass effect in the posterior fossa can lead to downward herniation of tonsils and subsequent obstruction of CSF flow at the foramen magnum.

Trapped Ventricles

The so called “trapped ventricle” may occur when the egress of CSF is obstructed, either from intrinsic or extrinsic masses. For example, in the trauma setting, a large subdural hematoma (SDH) can compress the ipsilateral lateral ventricle, but because of midline shift and outflow obstruction at the foramen of Monro, the contralateral lateral ventricle can abnormally dilate. Not uncommonly, periventricular hypoattenuation on CT or hyperintensity on T2/FLAIR around the margins of the trapped ventricle can be seen, indicating transependymal flow of CSF.

Trapping of the third ventricle is uncommon. Selective enlargement of the third ventricle must be distinguished from the presence of an intraventricular ependymal/arachnoid cyst and third ventricle squamopapillary craniopharyngiomas.

Isolation of the fourth ventricle may occur when the aqueduct of Sylvius, foramen of Magendie, and Luschka are occluded. The fourth ventricle becomes “trapped” and will expand as CSF production by the choroid plexus continues unabated ( Fig. 7-6 ). This expansion may compress the cerebellum and brain stem and lead to posterior fossa symptoms. Many of these cases are due to fibrous adhesions with or without earlier hemorrhage.

Infection, Hemorrhage, Tumors

The arachnoid villi are sensitive, delicate structures that may get gummed up by insults of several causes, resulting in communicating hydrocephalus ( Box 7-3 ). Think of them as the little fenestrations in your bathtub drain; the whole tub will overflow if these tiny conduits are obstructed. The most common causes of obstruction include infectious meningitis, ventriculitis, ependymitis, subarachnoid hemorrhage, and carcinomatous meningitis. As the CSF becomes more viscous with a higher protein concentration, the arachnoid villi lose their ability to reabsorb the fluid. This causes hydrocephalus with dilatation of the ventricular system.

Do not let a normal appearance to the fourth ventricle dissuade you from considering communicating hydrocephalus. The fourth ventricle is the last ventricle to dilate, possibly because of its relatively confined location in the posterior fossa, surrounded as it is by the thick calvarium and sturdy petrous bones. Thus it is not uncommon to see dilated lateral and third ventricles but a normal-sized fourth ventricle and have communicating hydrocephalus. Still the most sensitive indicator will be the enlargement of the temporal horns and/or anterior recesses of the third ventricle—without that you probably do not have hydrocephalus. The hunt for a source of the ventricular dilatation should not stop at the aqueductal level with this pattern.

As with any cause of hydrocephalus, there may be periventricular high signal intensity on MR, very nicely demonstrated with FLAIR scanning. This is due to transependymal CSF migration into the adjacent white matter leading to interstitial edema (dark on diffusion-weighted imaging [DWI]). This is most commonly seen at the angles of the lateral ventricles and, because of its smooth and diffuse nature, can usually be distinguished from the focal periventricular white matter abnormalities associated with atherosclerotic small vessel ischemic disease. Be aware that there may normally be mild high intensity at the angles of the ventricle (ependymitis granulosa) in middle-aged patients.

Adult Hydrocephalus

Normal-pressure hydrocephalus (NPH) or adult hydrocephalus has a classic triad of clinical findings; the recent onset of gait apraxia, dementia, and urinary incontinence ( Box 7-4 ). Half of patients have no known prior insult (idiopathic NPH), while the other half carry a remote history of prior infection or hemorrhage (nonidiopathic NPH). Imaging shows enlarged ventricles from communicating hydrocephalus with particular enlargement of the temporal horns. There may be evidence of transependymal CSF leakage on MR or CT. MR often shows accentuation of the cerebral aqueduct flow void ( Fig. 7-7 ), but this can be seen as a normal finding as well. These patients may respond to shunting or endoscopic third ventriculostomy procedures with amelioration of their clinical symptoms, making this a treatable cause of dementia (although the gait disturbance is more readily responsive to treatment). Because there is a chance at the possibility of return of function with a shunt, it is important to at least consider the diagnosis of NPH when ventricles appear larger than expected ( Fig. 7-8 ). The most accurate predictors of a positive response to shunting are (1) absence of central atrophy or ischemia, (2) gait apraxia as the dominant clinical symptom, (3) upward bowing of the corpus callosum with flattened gyri and ballooned third ventricular recesses, (4) prominent CSF flow void, and (5) a known history of intracranial infection or bleeding (nonidiopathic NPH).

In patients with suspected NPH, an indium 111-DTPA (diethylenepentaacetic acid) study is sometimes ordered. The agent is instilled in the CSF through a lumbar puncture. Normally, the tracer is resorbed over the convexities without ventricular reflux within 2 to 24 hours. In cases of communicating hydrocephalus and NPH, reflux of the tracer into the ventricles is seen with lack of tracer accumulation over the convexities 24 to 48 hours after instillation ( Fig. 7-9 ). Patients who demonstrate this scintigraphic appearance allegedly have a better response to shunting than patients with normal or equivocal indium findings.

The rate of clinical improvement after shunting of patients with NPH is still only 50%. Prominence of the CSF flow void in patients with this condition has led some investigators to use phase contrast MR techniques to measure the flow through the cerebral aqueduct. A stroke volume of greater than 42 mL has been shown to be predictive of better response to shunting. The specific parameters inherent in this measurement are related to scanner field strength and pulse sequences, so they are not necessarily transferrable to your own scanner, but the point made is that greater flow through the aqueduct means a better chance for shunt improvement.

The best predictor of therapeutic response to shunting seems to be the patient’s response to trials of large volume CSF drainages and/or a multiday trial of lumbar CSF drainage tube placement.

External Hydrocephalus

Another benign cause of hydrocephalus from arachnoid villi malfunction is “external hydrocephalus,” also referred to as “benign macrocephaly of infancy,” “benign enlargement of the subarachnoid spaces in infants (BESS),” “benign extraaxial collections of infancy,” “extraventricular obstructive hydrocephalus,” and “benign subdural effusions of infancy.” This may be due to immaturity of the arachnoid villi with a decreased capacity to absorb CSF. BESS is typically seen in children less than 2 years old who have a rapidly enlarging head circumference. Transient developmental delay may be present at the time of presentation; however, the clinical and imaging findings usually resolve by the time the child is 3 to 4 years old and the head circumference returns to normal. Prematurity, a history of intraventricular hemorrhage, and some genetic syndromes predispose to “external hydrocephalus.”

CT and MR show dilatation of the subarachnoid spaces over the cerebral convexities and normal or slightly enlarged ventricular system ( Fig. 7-10 ). The differential diagnosis includes chronic subdural hygromas and atrophy caused by previous injury. Sulcal dilatation and vessels coursing through the subarachnoid spaces indicate enlarged subarachnoid spaces, whereas displacement of vessels towards the surface of the brain imply the presence of a subdural collection. Atrophy is not usually associated with an enlarging head circumference. Beware, patients with BESS have an increased rate of SDHs presumably because the enlarged subarachnoid space causes stretching of the bridging veins in the subdural space. The conundrum is if nonaccidental trauma (NAT; now referred to as inflicted injury) is suspected in these patients because SDHs of different ages and volume loss may be present in NAT victims. Different aged SDHs would suggest NAT—along with additional findings of retinal hemorrhages and fractures (see Chapter 4 ).

The Almighty Failed Shunt

Shunt failure accounts for a large number of unenhanced CT scans in pediatric neuroradiology. The typical scenario is a child with a ventriculoperitoneal shunt in place who presents with nausea, vomiting, irritability and/or fever. This occurs in 30% of individuals in their first year with a shunt and in 50% of subjects within the first 6 years after shunt placement. Shunt infection occurs at a rate of about 10% in the first year.

Having prior imaging on hand is very valuable in the diagnosis of shunt failure. Evaluation should include a careful comparison of the entire ventricular system between the current exam and the prior, with search for interval change in size. Remember that there may be compartmentalization of the ventricular system in more complicated cases of brain tumors, ventricular hemorrhages, and infections, such that portions of the ventricular system may show interval change in size rather than the entire ventricular system. Next, contemplate the principal mechanisms responsible for shunt failure: (1) obstruction of the catheter tip in the ventricular system; (2) malfunction of the valve; (3) kinks in the tubing; (4) obstruction at the distal end of the catheter (i.e., intraperitoneal, intracardiac); and (5) component disconnection. The valves come in a variety of pressure settings for various resistances.

Shuntograms in which 2 to 3 mL of nonionic contrast are injected into the shunt reservoir may be revealing. Normally, the contrast clears from the shunt tube within 3 to 10 minutes. In adults, it may take 10 to 15 minutes to clear. The first step in the evaluation is withdrawing CSF from the shunt valve. If CSF cannot be withdrawn, the ventricular catheter is obstructed or the valve is faulty. If contrast refluxes from valve to ventricle, the valve is faulty as this is supposed to be a one-way valve to prevent “dreck” from the flowing backwards to the ventricles. If the contrast does not flow freely out but after pumping the valve it seems to work, there is probably incomplete obstruction of the shunt system and/or a malfunctioning valve-pressure system. If there is no spillage intraperitoneally even after pumping the valve, or if what spills gets loculated, clearly, there is a problem with the end of the shunt system. Ventricular catheter obstruction, valve malfunction, and distal obstruction are the most commonly seen causes of shunt failure.

Third ventriculostomies, where a small hole is made at the floor of the third ventricle that allows communication between the floor of the third ventricle and the suprasellar cistern, has proven to be effective in relieving hydrocephalus (see Fig. 7-4 ). This is most useful for those obstructions distal to the third ventricle as it bypasses the obstructed region. These are often placed through the use of a fiberoptic endoscope and/or three-dimensional reconstructions with image-guided navigation. Expect the reduction in ventricular size to appear within a couple of weeks of the procedure—not as rapidly as with lateral ventricular shunts. Flow through the third ventriculostomy may be visualized with phase contrast flow studies at a velocity encoded at 5 mL/min.

At those institutions who have ready access to MR, radiologists are encouraging the use of single shot fast spin echo MRI (timing in at 1 to 2 seconds per slice) for evaluation of shunted kids to avoid the repeated irradiation of children’s heads with CT. No more frying of the brain with CT for suspected shunt failure—that’s a good thing!