gms | German Medical Science

Objective: The evaluation of a suspected malfunction of a ventricular shunt is one of the most common procedures in neurosurgery. In practice, the evaluation rely either on the interpretation of the ventricular width using static radiographic images (e.g. cranial CT or MRI) or on invasive techniques (e.g. surgical revision or radionuclide study). During the last decades several attempts have been made to measure the flow velocity of the cerebrospinal fluid (CSF) utilizing different MRI techniques. Particularly the signal to noise ratio of the evaluated 1,5 Tesla MRI scanners and the increasing number of adjustable, magnet containing valves are limiting a more common use. In the present study we evaluated the value of 3-Tesla MRI scanners for the determination of the low flow in ventricular shunt systems.

Methods: All MR-imaging procedures were conducted using a Siemens Prisma scanner with 3-Tesla magnet field strength. The flow was measured with phase-contrast-sequences that were modified to work properly in the expected range of flow rates. As a first step, an MRI-phantom of the brain consisting of a mixture of fatty acids to mimic a brain was used to measure the phase-contrasts at different constant flow rates. The next step was to measure the CSF flow in patients treated with ventricular shunts who were at the time point of the measurement in clinical stable setting without any suspects of a malfunction of the observed shunt.

Results: The phase-contrasts of the phantom showed a linear correlation between the CSF flow and the depending signal intensities. Under these ideal conditions flow rates down to 0,5 ml per hour are reliably quantifiable. First measurements of the flow in patients with shunted hydrocephalus confirm these results. Despite a lot of artifacts due to the valve flow signals of the shunts were measurable within the parenchymal portion of the shunt in all patients. However, the signal intensities of the phase-contrast MR images depend on the underlying cause of hydrocephalus. Thus, the flow rate was always detectable in patients with obstructive hydrocephalus while in patients with normal pressure hydrocephalus the detected signal intensities were close to the limit of measuring.

Conclusion: CSF flow detected within the parenchymal portion of the shunt by phase-contrast MRI may reliably provide information about the flow rate of the ventricular shunt. Even in patients whose hydrocephalus were treated with adjustable valves the CSF flow is detectable using 3-Tesla phase-contrast MRI sequences.