Flow-sensitive MR imaging of ventriculoperitoneal shunts: in vitro findings, clinical applications, and pitfalls.

To determine ventriculoperitoneal shunt patency during routine MR imaging of the head, 23 patients were studied with T1-weighted fast-field-echo scans. Without knowledge of the results of previous MR/CT studies or of the patients' clinical history, we reviewed the fast-field-echo studies and divided them according to those judged to have shunt flow (18) and those judged not to have flow (five). Fast-field-echo sequences showed high signal intensity, consistent with CSF flow and shunt patency in 17 medium-pressure systems and one high-pressure system. No signal was seen in five patients with high-pressure valve shunts. Combined clinical evaluation and MR/CT studies showed that three patients had probable shunt malfunction. One patient had true shunt malfunction; and although malfunction was thought to be present in two symptomatic patients, surgical revision showed the shunts to be patent. The possibility of temporary shunt obstruction is postulated to explain the clinical and MR findings in those two cases. The remaining two cases (9% of the patients) had no clinical evidence of shunt malfunction, and the MR findings probably reflected periodic CSF flow. One patient had an intracranial segment that was not connected and showed no flow on MR. No false-positive results (apparent flow in a nonfunctioning shunt) occurred. Using a standard medium-pressure shunt system, we constructed and imaged a phantom, which confirmed our clinical observations. T1-weighted fast-field-echo sequences may be useful in assessing patency of medium-pressure CSF shunt systems.