11 – Core imaging in adult hydrocephalus

11 Core imaging in adult hydrocephalus

Adult Hydrocephalus, ed. Daniele Rigamonti. Published by Cambridge University Press. © Cambridge University Press 2014.

Benjamin D. Elder

Patrick Bankah

Ari M. Blitz

C. Rory Goodwin

Thomas A. Kosztowski

Ignacio Jusué-Torres

Daniele Rigamonti

Introduction

Advances in neuroimaging have greatly improved the management of hydrocephalus. The various imaging modalities are useful in confirming the diagnosis, determining the etiology of hydrocephalus, providing intraoperative guidance, and improving the follow-up of patients who have undergone treatment including cerebrospinal fluid (CSF) shunting or endoscopic third ventriculostomy (ETV). The various imaging modalities that are utilized include:

conventional X-rays;
ultrasonography;
computerized axial tomography (CT);
magnetic resonance imaging (MRI);
radionuclide shunt patency imaging.

As systematically as possible, we will list for each modality their use in the preoperative, perioperative, and postoperative period.

Conventional X-rays

Conventional X-rays often are used to evaluate shunted patients as the entire shunt hardware can be visualized to assess the continuity of the shunt. However, as the majority of shunt malfunctions are unrelated to mechanical discontinuity, the majority of images obtained will demonstrate intact tubing; Lehnert et al. [1] found only 4% of shunt surveys to be abnormal in a series of 296 patients in the emergency room. Similarly, Griffey et al. [2] assessed the utility of shunt surveys in diagnosing shunt malfunctions in the emergency room in adult hydrocephalus patients and found a sensitivity of only 0.11 but a specificity of 0.98. A similar study in the pediatric population demonstrated a slightly higher sensitivity of 0.31 [3]. Conventional X-rays are also utilized to verify the shunt setting, because programmable shunt valves generally have radiopaque markers.

In addition to the use of plain films for diagnosing shunt discontinuities, we routinely obtain plain-film shunt surveys immediately postoperatively as a baseline and to verify shunt continuity. For pleural or atrial distal catheters, we also obtain a chest X-ray (often as part of the shunt survey) to verify the location of the distal catheter and to confirm that there is no pneumothorax (Figure 11.1). For peritoneal distal catheters, we perform an abdominal X-ray 12–24 hours after the shunt survey to confirm movement of the distal catheter as evidence that it is not pre-peritoneal or in a loculated cavity (Figures 11.2 and 11.3).

Figure 11.1 A postoperative chest radiogram shows correct and incorrect placement of the distal tip of the ventriculoatrial catheter.

Figure 11.2 A postoperative abdominal radiogram shows that the tip of the catheter is moving in the course of several hours because of the peristalsis, confirming intra-abdominal placement.

Figure 11.3 A postoperative abdominal radiogram and a later one show that the tip of the catheter, initially in the abdomen, migrated to an extraperitoneal subcutaneous location.

Ultrasonography

The use of duplex ultrasound of the internal jugular vein is often helpful in preoperative planning for ventriculoatrial shunt placement to confirm patency of the vein; it is being used intraoperatively to guide the percutaneous insertion of the distal catheter (Figure 11.4) and even to navigate around venous obstructions [4].

Figure 11.4 Ultrasound visualizes a pocket of CSF surrounding a proximal lumbar catheter that migrated from the spinal canal to the subcutaneous location.

Ultrasonography is useful in the assessment of possible fluid collections around the distal shunt tubing in peritoneal, pleural, and atrial (Figure 11.5) catheters that may be suggestive of shunt malfunction or infection. However, a CT scan of the chest or abdomen is required if more detailed imaging is needed to better assess the fluid collection.

Figure 11.5 Thoracic ultrasound visualizes a small thrombus at the distal tip of a ventriculoatrial shunt.

CT imaging

Diagnosis of hydrocephalus

CT scans are often the first-line imaging in hydrocephalus due to wide availability, ease with which they are performed, minimal contraindications, and low expense relative to MRIs. Generally, in the presence of hydrocephalus, the temporal horns, which are normally barely visible, look dilated. The appearance of the sylvian fissure, interhemispheric fissure, and cerebral sulci depends on the type of hydrocephalus. They are barely visible in the case of obstructive hydrocephalus and clearly visible in idiopathic normal pressure hydrocephalus (iNPH). An Evans ratio (defined as the ratio of the frontal horn to the maximal biparietal distance measured in the same CT slice) greater than 0.3 is generally considered a reliable sign of hydrocephalus (Figure 11.6) [5]. The Japanese guidelines for iNPH also describe this measurement [6], although some groups advocate the use of volumetric ventricular measurements in the diagnosis of iNPH now that high resolution CT scans are widely available; one recent study proposed a ventricular volume greater than 77 ml as suggestive of iNPH [7]. Hydrocephalus is also suggested on imaging when the size of both temporal horns (TH) exceeds 2 mm in width. Additionally, a ratio of the FH/ID exceeding 0.5 is suggestive, where FH is the largest width of the frontal horns and ID is the internal diameter from inner table to inner table at the same level. Other generally accepted evidence of hydrocephalus includes ballooning of the frontal horns of the lateral ventricle (“Mickey Mouse” ventricles), or transependymal flow suggestive of acute hydrocephalus.

Figure 11.6 Different from the original description based on coronal pneumo-encephalography, we routinely measure the Evans index as the ratio between the maximum frontal horn span, measured in front of the caudate nucleus, and the maximum biparietal inner skull diameter in the same section.

Volumetric analysis (VA) by CT or MRI of the CSF spaces

VA is an interesting tool to assess the size of the ventricles, the size of the subarachnoid space (SAS), their relationship, and changes in the size of the ventricles or SAS after shunting [8]. Although it seemed to hold great promise, the tool is presently considered still experimental and its clinical use is limited (Figure 11.7).

Figure 11.7 The preoperative and postoperative ventricle volumes are measured using non-enhanced 1.5–3 mm CT sections. The sum of areas from all axial sections is used to compute ventricle volume using MIPAV software in a semi-automated fashion (OsiriX, Los Angeles, CA Center for Information Technology, National Institutes of Health, Bethesda, MD).

Postoperative management: shunt malfunction

CT scans are performed to obtain a baseline image postoperatively to demonstrate the location of the proximal catheter in the ventricular system, the baseline size of the ventricles, and any postoperative intraventricular hemorrhage (Figure 11.8). CTs are helpful in diagnosing shunt malfunctions in younger patients: a CT image should ideally be compared to a prior baseline image obtained when the patient was doing well clinically; however, this is not always possible, particularly in the emergency room setting, though every attempt should be made to obtain prior imaging with which to compare. In general, an increase in ventricular caliber may be indicative of a shunt malfunction, or may be an expected outcome if a programmable shunt was recently changed to a higher setting. Likewise, a decrease in ventricular caliber may be due to a recent programmable valve change to a lower setting or recent shunt revision due to malfunction, but can be a sign of overdrainage.

Figure 11.8 Early postoperative CT shows the new intracerebral hemorrhage (ICH) secondary to the insertion of a ventriculoperitoneal shunt.

A recent retrospective study examined the utility of head CT scans in a series of 296 patients who presented to the emergency room, and found that 18% of the scans were positive and that there was a significant association between abnormal head CT and surgical revision [1]. Another study demonstrated the importance of obtaining prior imaging for comparison as having increased ventricular size since a prior CT scan had a sensitivity of 48% for shunt malfunction, but enlarged ventricular size without a prior comparison CT only had a sensitivity of 28% for shunt malfunction [9]. Finally, particularly with bone windows, CT images can verify the continuity of the tubing and valve connection in the cranial region. However, in adult patients shunt malfunction is less likely to be diagnosed on the basis of CT, because in these situations there may be minimal change in ventricular size. Patients with iNPH [10] or pseudotumor cerebri, or patients with the slit ventricle syndrome or decreased compliance [11], may not demonstrate a change in ventricular size on CT imaging. CT imaging is also ideal to rule out subdural hematoma or hygroma formation in the setting of overdrainage (Figure 11.9).

Figure 11.9 CT shows a left subdural hygroma and its disappearance after adjusting the valve to a higher setting.

In addition to its use in diagnosing shunt malfunction with cranial imaging, CT scanning is very helpful in diagnosing abdominal shunt-related complications in patients with peritoneal distal catheters. For example, Chung et al. [12] examined a series of 70 patients with ventriculoperitoneal shunts who presented with abdominal complaints and found that almost 23% of the patients had evidence of abdominal shunt-related complications including shunt migration (Figure 11.10), infection, CSF pseudocyst, abdominal abscess, infected fluid collection, or rarely bowel or bladder perforation (Figure 11.11).

Figure 11.10 Chest CT confirms the migration into a subcutaneous location of a ventriculopleural shunt.

Figure 11.11 Abdominal CT confirms the perforation into the bladder of a ventriculoperitoneal shunt.

Despite its widespread use in hydrocephalus diagnosis and treatment, CT scanning is not without its risks, as the patient is exposed to a large dose of ionizing radiation with each scan. This is especially a drawback in the pediatric population. For instance, a recent review estimated a 1% lifetime risk of head and neck malignancies from diagnostic radiation exposure in children with ventriculoperitoneal shunts [13]. To mitigate this risk, many centers have developed lower-dose head CT protocols for both the adult [14] and pediatric [15] populations.

MRI

MRI is considered the gold standard imaging modality in the evaluation of hydrocephalus in adults. In addition to providing much of the same diagnostic evidence and measurements as can be obtained on CT imaging, MRI can provide substantially more anatomic detail. It provides excellent multiplanar anatomic definition of the ventricular system, subarachnoid spaces, and brain parenchyma. It is the preferred examination for the diagnosis of iNPH, especially with T2-weighted images. T2 steady-state free precession (SSFP) images (e.g. CISS or FIESTA) are extremely valuable in outlining the fine anatomy of the aqueduct, third ventricle, foramen of Monro, fourth ventricle, and the subarachnoid space. This sequence can confirm the patency of the aqueduct prior to a large volume lumbar puncture or lumbar drainage (Figure 11.12a and b). The sequence can outline an obstruction due to a web (Figure 11.12c) or tumor at the level of the aqueduct or of the foramen of Monro. In the presence of an obstruction, the walls of the third ventricle are characteristically bulging outward (Figure 11.12c). The same sequence is used to document the patency of a third ventriculostomy after ETV (Figure 11.12d). Coronal MRI sequences may demonstrate a disproportionately enlarged subarachnoid space hydrocephalus (DESH) pattern that the Japanese authors consider pathognomonic of idiopathic NPH (Figure 11.13) [6]. Finally, transependymal CSF flow suggestive of acute hydrocephalus may be seen on T2 or FLAIR imaging (Figure 11.14).

Figure 11.13 A coronal T1-weighted MRI demonstrates crowding of the sulci over the convexity and the enlarged subarachnoid space of the sylvian fissure that produce the characteristic DESH pattern described by the Japanese authors [6].

Figure 11.14 T2-weighted and FLAIR MR images show the characteristic “capping” of the frontal horns secondary to presumed retrograde CSF reabsorption into the parenchyma.

Figure 11.12 (a) Sagittal midline CISS image demonstrates anatomic patency of the cerebral aqueduct (arrow). (b) Aqueductal patency is confirmed with the use of cardiac gated phase-contrast imaging. Hyperintensity on the magnitude of phase images demonstrates the presence of CSF flow in the cerebral aqueduct (arrow). (c) Sagittal midline CISS images in another patient demonstrate obstruction of the caudal aspect of the cerebral aqueduct due to two thin webs (arrow). CSF flow could not be demonstrated through the region of the cerebral aqueduct on cardiac gated phase-contrast images (not shown). (d) The patient underwent endoscopic third ventriculostomy, the patency of which is confirmed by an anatomic defect in the floor of the third ventricle on CISS images through which a flow void extends (arrow). Cardiac gated phase-contrast images confirmed the presence of CSF flow (not shown).

MRI does not require ionizing radiation, making it ideal for use in the pediatric population or when a patient requires serial imaging. However, some drawbacks to MRI use include the time length of the scan, need for sedation or anesthesia in patients with claustrophobia or pediatric patients, multiple contraindications due to metal implants, difficulty performing scans in patients with pacemakers or implanted defibrillators, as well as the significant expense relative to CT scans. It should also be noted that patients with older programmable shunts must have the valve reset following the MRI as the older valves can have their setting changed by the magnetic field during the scan.

Phase-contrast MRI

Phase-contrast or “cine” MR imaging was originally applied to CSF imaging by Levy and Di Chiro [16] in order to demonstrate pulsatile CSF flow, and the process was later refined to better quantify CSF flow rates [17]. It has been most useful in the diagnosis of aqueductal stenosis [18] and CSF obstruction, as described above, and has been shown to be more sensitive than conventional MR imaging (Figure 11.12b). For instance, in a retrospective study, 17/17 patients were found to have altered flow with aqueductal obstruction on phase-contrast imaging whereas only 10/17 were found to have aqueductal stenosis on conventional MR imaging [19]. It can also be used to determine the patency of a prior ETV [20]. Phase-contrast imaging also has some utility in identifying proximal shunt flow to assess for proximal malfunction. Although it has proven beneficial in the diagnosis of aqueductal stenosis and proximal shunt flow, a prior study demonstrated that it was not beneficial in predicting which iNPH patients will respond to shunting [21], though research is still ongoing in this area.

MRI with gadolinium enhancement

The administration of gadolinium allows the visualization of vascular structures. The appearance of a thickened dural layer suggests the presence of intracranial hypotension secondary to overdrainage of CSF by the shunt (Figure 11.15).

Figure 11.15 On gadolinium-enhanced T1-weighted images the appearance of diffuse pachymeningeal enhancement, even in the absence of brainstem sagging, is suggestive of intracranial hypotension.

Radionuclide shunt patency test

It is not always easy to diagnose shunt obstruction. The presence of enlarging ventricles, compared to historical baseline, on standard imaging modalities such as CT or MRI is usually sufficient proof. However, even with enlarging ventricles on cranial imaging, it is difficult to assess if malfunction is at the level of the proximal catheter, valve, or distal catheter. The problem is compounded in iNPH as shunting results in a minimal change in ventricle size [22], and there is a similarly minor change in ventricle size with shunt malfunction. A similar diagnostic dilemma can be encountered when assessing possible shunt malfunction in children with slit ventricle syndrome or pseudotumor cerebri patients.

We have employed the use of radionuclide shunt patency (SP) studies in the work-up of shunt malfunctions in these groups. This procedure was developed in 1966 by Di Chiro and Grove [23], and has been modified by other groups with water-soluble contrast medium [24], termed a “shuntogram.” The radionuclide SP study was recently studied and validated in the iNPH population by our own group [25].

The SP study allows for the assessment of four variables in the diagnosis of shunt malfunction which also isolate the region of the shunt that has malfunctioned. These include ventricular entry, opening pressure, T_1/2 time, and peritoneal distribution. Using all four variables, the positive predictive value of the SP study was 80% [25]. Ventricular entry of radiotracer helps rule out proximal malfunction. Elevated opening pressure with brisk flow of CSF indicates malfunction of the valve or distal catheter, or demonstrates an inappropriate shunt setting in a programmable system. The peritoneal distribution of radionuclide is classified as diffuse, loculated, or none, and indicates potential failure of the distal shunt tubing.

The T_1/2 time is calculated as the time for radioactivity counts in the shunt bulb to decrease to half of the starting value (Figure 11.16). In iNPH patients, a T_1/2 >8.7 min is suggestive of shunt malfunction, with a positive predictive value of 87.5% [25]. However, the T_1/2 is less accurate in patients with low CSF pressures (<5 cmH₂O), as there may be no spontaneous CSF flow during the SP study in the supine position, when the T_1/2 is calculated.