Material and Method: The retrospective study was performed in patients suspected NPH and underwent MRI CSF flow measurement. 2D-phase contrast technique (Achieva, 3 Tesla Philips system) was used as CSF flow analysis. The preoperative and postoperative clinical outcomes were collected and analyzed to determine predictive value of MRI CSF flow measurement in shunt responsive patients.

Results: Between 2006 and 2011, twenty NPH patients underwent MRI CSF flow study and were treated by ventriculoperitoneal shunt placement. Fourteen of 20 cases had improved, at least in gait score. Of these, 10 were defined as significant responsive group for overall improvement of outcome (sum of iNPHGS ≥ 3). The mean velocity of the CSF flow through the aqueduct of Sylvius was significant difference between shunt-responsive and non- responsive groups (p < 0.05). The peak velocity was a significant difference between gait responsive and non-responsive groups (p < 0.05). Using a mean velocity threshold 26 mm/sec to identify the significant responsive group, the sensitivity is 50%, specificity 83.3%, positive predictive value 87.5%, and accuracy 70%. In order to identify the gait responsive group by using a threshold of peak velocity 70 mm/sec, the sensitivity was 60%, specificity 83.3%, positive predictive value 81.5%, and accuracy 60%.

Conclusion: Using available commercial software in the authors’ institute, the mean velocity as well as the peak velocity was a specific value that predicted significant shunt responsiveness in NPH patients.

Keywords: Normal pressure hydrocephalus (NPH), MRI CSF flow study, Idiopathic NPH grading scale (iNPHGS)