T1- and T2-weighted are the sequences routinely used. Both are quite informative for demonstrating pathologic signal intensities within the brain parenchyma and for the assessment of ventricular shape and size. In addition, these sequences show most of the space-occupying lesions in both intra- and extra-axial locations (Fig. 3.4) [8]. Conventional MRI also provides useful information about the etiology of the hydrocephalus. Nevertheless, these criteria depend on a subjective evaluation by the neuroradiologist that may render the images difficult to assess in some instances, and thus they may be hardly comparable in postsurgical studies [17, 18]. The demonstration of the presence or absence of hemorrhage and of related membranes in the cisterns by using hemosiderin-sensitive sequences under the age of 2 years is also essential. Gradient echo (GRE) T2* or susceptibility weighted imaging (SWI) easily detects previous hemorrhage in the ventricles and cisterns.

Three-dimensional constructive interference in the steady state (3D-CISS) is a gradient-echo imaging technique with high resolution that remains sensitive to flow and provides fine anatomic details of the CSF pathways. 3D-CISS enables locating the obstruction and determining the upstream impact. It provides anatomical information about third ventricle morphology and its relationships, details that are extremely useful before performing an ETV. These sequences also allow for a good visualization of the cerebral aqueduct and for the diagnosis of the cause of obstruction, sometimes better than with classical sequences [8, 15, 16, 19, 20]. The value of 3D-CISS has been already demonstrated in hydrocephalus, and its advantage has been shown not only for demonstrating anatomical details but also for depicting the presence, location, and extension of membranes within the cisterns, which are of help for guiding ETV [16, 21]. Membranes in the foramen of Monro, superior velum medullaris, fourth ventricle outlets, and even intraventricular cystic masses may not visible in conventional T1- and T2-weighted images but are depicted with these sequences [16]. We routinely obtain sagittal 3D-CISS images in the midsagittal plane covering midline structures, basal cisterns, and fourth ventricle outlets with an extremely high resolution that allows reconstructions in any plane without losing data (Fig. 3.5).

Cardiac-gated cine phase-contrast (PC) MRI is the only technique currently available to noninvasively detect CSF flow. It is a rapid, simple, and noninvasive technique that is sensitive to CSF flow [8, 15, 22, 23]. The PC-MRI generates signal contrast between flowing and stationary nuclei by sensitizing the phase of the transverse magnetization to the velocity of motion [24]. The technique provides “to-and-fro” CSF flow velocity and direction during a single cardiac cycle with a specially designed flow-sensitive GRE sequence. CSF flow is pulsatile and synchronous with the cardiac cycle; therefore, cardiac gating can be used to provide increased sensitivity [24, 25]. The cardiac-gated phase-sensitive technique is based on the subtraction of two similar GRE images, one with flow-encoding activated gradients and another with identical parameters but without these activated gradients. The difference in phase between the subtracted images is due to motion along the particular gradient axis. In this way, the signal from the stationary tissues is almost completely eliminated. Two series of PC imaging techniques are applied in the evaluation of CSF flow, one in the axial plane, with through-plane velocity encoding in craniocaudal direction for flow quantification, and another in the sagittal plane with in-plane velocity encoding in the craniocaudal direction for qualitative assessment [24].

The sensitivity of the flow-encoding gradients must be prospectively set to prevent aliasing. Qualitative midline sagittal and quantitative axial images perpendicular to the imaging plane, mostly the aqueduct, must be acquired. It takes a maximum of 10 min to get both sequences with gating [15, 16]. Flow in craniocaudal direction is encoded in shades of white, while flow in caudocranial direction is encoded in shades of black.

CSF flow MRI can be used to distinguish communicating and non-communicating hydrocephalus, to localize the level of obstruction in obstructive hydrocephalus, to detect communication of arachnoid cysts with the subarachnoid space, to differentiate between arachnoid pouches and subarachnoid space, to discriminate between syringomyelia and cystic myelomalacia, and to evaluate flow patterns in posterior fossa cystic malformations. This imaging method can also provide significant information in the preoperative evaluation of the Chiari 1 anomaly and of normal pressure hydrocephalus (NPH) and also in the postoperative follow-up of patients given an ETV or a VP shunt [24].

But cine PC has several disadvantages. To prevent aliasing artifact, which could affect the quality of both qualitative and quantitative assessments of cine PC, the strength of the velocity-encoding gradient needs to be set properly before scanning. On the other hand, cine PC demonstrates only bidirectional flow in a selected direction [16]. Whereas cine PC shows CSF flow, it is unable to demonstrate cisternal anatomic details and should be used in conjunction with 3D-CISS. Furthermore, this technique is limited because of poor visualization of turbulent flow and of its inability to measure bulk flow [8, 23].

Hydrocephalus can be treated by CSF diversion or by endoscopic third ETV [26]. Ventricular shunting represents the traditional treatment for hydrocephalus. The anatomy of a ventricular shunt consists of three parts: the proximal catheter, the valve (with or without a reservoir), and the distal catheter. The catheters are typically made of silicone and have a diameter of 2–3 mm. Catheters are hypointense on T1- and T2-weighted MRI images and hyperdense on CT. The reservoir and valve are normally radiolucent, but they have radiopaque markers to allow for visualization (Fig. 3.1). The tip of the proximal catheter is placed at the widest part of frontal horn of the lateral ventricle, away from the choroid plexus and in front of the foramen of Monro [5]. The proximal catheter may be inserted either into the frontal or occipital regions through a small burr hole in the skull [26]. The catheter exits via the burr hole and connects to a reservoir that allows CSF fluid sampling and intraventricular pressure measurements. The reservoir is connected to the inflow port of a one-way valve, while the distal catheter is joined to the outflow port of the valve and is tunneled subcutaneously to end in the drainage cavity. The valve regulates the amount of CSF drainage and plays an essential role in the successful treatment of hydrocephalus [5]. In programmable shunts, the valve contains magnets that enable transcutaneous adjustments of the valve’s opening pressure [2, 4]. Shunts are commonly placed in the peritoneum, right atrium, or pleural space. VP shunting is the preferred method for CSF shunting by most neurosurgeons due to fewer complications and to the relative ease of peritoneal access [27]. Intra-abdominal pathology, such as peritoneal adhesions or recurrent peritonitis, necessitates placement or relocation into an alternative drainage site [4]. VA shunts are generally chosen when VP shunting has already failed or is contraindicated. The right atrium may be accessed percutaneously via the facial, subclavian, or internal jugular vein [28], and proper placement of the distal VA catheter may be ensured by real-time transesophageal echocardiography. The primary disadvantage of VA shunts is their risk of serious intravascular complications. On the other hand, ventriculopleural shunts have been rarely used for long-term shunting because of the high incidence of hydrothorax. Some authors have reported using ventriculopleural shunts for temporary CSF drainage in cases of VP or VA shunt infection and of tumoral hydrocephalus [4].

Radiologists are usually asked to assess shunt function by determining the position of the shunt tubing and for estimating the ventricular size. Decrease in ventricular size after shunting depends on several factors among them the cause of hydrocephalus. A prompt reduction as shown by a 24-h after shunting CT is expected in uncomplicated cases. CT scan will depict a normal or slit-like appearance of the ventricles in 60–80 % of shunted children [5]. If the ventricular size remains unchanged after 1–3 days, the patency of the shunt system should be questioned [26].

ETV is an increasingly popular treatment of hydrocephalus since it avoids many of shunt-related complications, especially those of shunt dependence and overdrainage [21]. The main indication for performing an ETV is obstructive hydrocephalus, for example, in cases of aqueduct stenosis or tumors at its vicinity [26]. Age is the main determinant of outcome after ETV, with younger children, especially neonates, having a poorer prognosis [29]. Specifically, ETV resolves hydrocephalus in 80 % of children older than 2 years of age, but is not as effective in children younger than 2 years due to immaturity of the arachnoid granulations [21, 30, 31]. In ETV, an endoscope is used to create a hole in the floor of the third ventricle, immediately anterior to the mammillary bodies. It allows the passage of intraventricular CSF to the interpeduncular and suprasellar cisterns, from where it can flow laterally and upwards over the cerebral convexities to sites of CSF resorption [26]. If ETV does not result in adequate resolution of hydrocephalus, a ventricular shunt can still be placed.

Endoscopic management must also be considered before shunt replacement in cases of CSF shunt malfunction to avoid shunt-related complications. Some cases require staged endoscopic procedures to adequately communicate cavities in loculated hydrocephalus [32]. Decrease in ventricular size following ETV, in contrast with shunting, may be more gradual and occurs over several weeks [31]. MR is the study of choice for the assessment of the success of an ETV (Fig. 3.5) as only MR can assess the patency of the surgically created stoma in the floor of the third ventricle and observe the CSF pulsating through it [26].

Intraventricular hemorrhage during shunt revision in shunted patients is a potentially serious complication due to the almost constant adhesions of the proximal catheter tip to the choroid plexus and/or to ventricular walls [5, 33, 40]. Hemorrhage complicating shunt failure is a known predictor of future ETV failure. Intraventricular bleeding arises from subependymal vessels or from the choroid plexus when these vascular structures are injured by the ventriculoscope or by the endoscopic instruments. Generally, the blood loss is stopped by means of a continuous irrigation with Ringer’s solution to clear the clots or even one has to resort to placing an EVD if the hemorrhage continues [5, 33–35]. Postoperative neuroimaging may demonstrate a subependymal hematoma (Fig. 3.6) that might require an adjunctive surgical procedure for its management [33, 36].

The reported incidence of intraventricular hemorrhage ranges around 1–3 % of the cases [34, 37], although some authors indicate a higher incidence ranging between 3.5 and 6 % in infants [38, 39]. This higher rate could be explained in part by taking into account the less favorable ventricular anatomy in this age group, such as smaller Monro’s foramina and larger interthalamic adhesions, which would increase the risk of surgical damage during the surgical procedure [33].

Hemorrhages within the ventricles, which many times evolve in an urgent scenery, are usually assessed by CT.

The occurrence of intracerebral hematoma due to damage to the brain vessels during the introduction of the proximal catheter or the ventriculoscope, although possible, is rather uncommon but possesses an important potential morbidity [5, 33, 35, 41]. Large intraparenchymal and intraventricular hemorrhages are disastrous complications, especially in patients who have coagulopathies or are receiving anticoagulants (Fig. 3.6) [2]. These hemorrhages are well delineated in CT studies, and the smallest one does not require further surgical treatment.

Subdural hematomas or hygromas usually follow a benign course [5, 42]. In acute subdural hematomas, excessive loss of CSF during the endoscopic procedure is probably the main pathogenetic mechanism [33]. Post-shunting subdural collections have been reported to be more frequent in patients with long-standing macrocrania and subsequent craniocerebral disproportion characterized by a remarkably decreased intracranial compliance [33, 43]. On the contrary, subdural fluid collections after EVT have been only occasionally described (an overall rate of 0.5–1.5 % was reported in a large series) [35].

Neuroimaging studies (CT or MRI) contribute to the diagnosis of these subdural collections and are used to follow-up their evolution and to monitor the results of treatments. In general, the appearance of these collections does not differ from the normal images of subdural hygromas and hematomas found in other conditions.

The presence of intracranial air following CSF shunting or endoscopic procedures is usually unimportant. However, tension pneumocephalus represents infrequent but dangerous complication [44]. Intracranial air (Fig. 3.6) results from the reduction in volume of the brain due to perioperative CSF leaks during the extrathecal CSF shunt implant or during the procedure of ETV [33, 45]. The presence of pneumocephalus, its volume, and distribution is clearly depicted with the current neuroimaging procedures (CT, MRI, and even with plain skull radiographs).

Another extremely rare complication of CSF overdrainage is the appearance of ventricular collapse and associated cerebral mantle inversion that are clearly depicted by CT or MRI studies [96]. They may require shunt replacement with flow-regulated valves, valve pressure upgrading, or addition of siphoning retarding [5]. Another unusual complication of CSF shunting is the occurrence of subarachnoid hemorrhage that is also well delineated by CT or MRI studies.