A prospective comparison between three-dimensional magnetic resonance imaging and ventriculography for...

The purpose of this prospective study was to compare stereotactic coordinates obtained with ventriculography with coordinates derived from stereotactic computer-reconstructed three-dimensional magnetic resonance (3D-MR) imaging in functional stereotactic procedures.

In 15 consecutive patients undergoing functional stereotactic procedures, both preoperative frame-based stereotactic 3D-MR imaging and intraoperative ventriculography were performed. Differences between 3D-MR imaging and ventriculography in X, Y, and Z coordinates of the anterior commissure (AC), posterior commissure (PC), and target area were calculated, as well as the 3D distance between the position of AC, PC, and target within stereotactic space as obtained using both methods. The position of the stereotactic MR imaging fiducial markers measured using 3D-MR imaging compared well with the markers' known position embedded in the software (mean error 0.4 mm, maximal error for an individual slice 1.2 mm). For the individual coordinates, only for Y-PC was a difference found between 3D-MR imaging and ventriculography that significantly exceeded half the size of a pixel, the theoretical limit of precision when using a digitized imaging technique. However, the mean difference was smaller than 1 mm. The mean 3D distance between the 3D-MR imaging- and ventriculography-derived coordinates was 1.09 mm for AC, 1.13 mm for PC, and 1.29 mm for the targets.

With these data it is shown that there is sufficient agreement between ventriculography-derived and 3D-MR imaging-derived stereotactic coordinates to justify the use of 3D-MR imaging target determination in frame-based functional stereotactic neurosurgery.