Determination of Programmable Shunt Setting Using CT: Feasibility Study

Programmable shunts can be adjusted to optimize CSF diversion in patients with hydrocephalus without the need for re-operation. Currently, all shunts incorporate radiopaque markers so that their setting can be determined on skull X-ray images. The purpose of this study was to evaluate whether the shunt setting could also be determined ex vivo and in vivo using the data from a standard head CT scan since one is nearly always obtained when patients with VP shunts present with new symptoms that could be due to shunt malfunction.  Materials and Methods: Four commonly used programmable shunts were attached to a dried skull and scanned using a variety of CT techniques. The shunts imaged were the Certas^TM Plus (Codman, Raynham, Massachusetts), Polaris^® (Sophysa, Orsay, France), proGAV 2.0^® (Braun, Bethlehem, Pennsylvania), and Hakim^® (Codman, Raynham, Massachusetts). Each shunt was scanned at two different valve settings using multiple CT techniques: CTDI_vol 75, 140kVp, 330mAs, CTDI_vol60, 120kVp 390mAs, CTDI_vol40, 80kVp with 430mAs, 140kVp with 215mAs. Image reconstruction with and without CT metal suppression software was used for all scans, and the data was reconstructed into volume-rendered images. We enlisted ten observers to review the volume-rendered images only. After a short set of training slides viewed by all observers, they were asked to predict the shunt setting for each valve along with their level of confidence. One clinical case of a patient with a programmable valve was evaluated on a CT scan.

Using the volume-rendered images only, the two shunt settings of the Polaris shunt were correctly predicted by all the observers and in nine of 10 settings for the Certas^TM Plus valve. For the Hakim^® shunt and the proGAV 2.0^® shunt, setting prediction accuracy was 0% and 10%, respectively. In one clinical case, the programmable valve setting could be determined from the CT scan data.

The valve setting of at least two currently available programmable shunts can be determined using volume-rendered images generated from CT data. Reconstructions using metal suppression software were rated as superior and may be necessary for some valve designs.