Magnetic resonance imaging versus computed tomography for target localization in functional stereotactic neurosurgery

To determine whether magnetic resonance imaging (MRI), compared with computed tomography (CT), provides consistent and accurate target localization for ventrolateral thalamotomy and posteroventral pallidotomy.

For 93 procedures (78 pallidotomies and 15 thalamotomies) in 83 patients, coordinates for the anterior commissure, posterior commissure, and stereotactic target were calculated from MRI- and CT-derived images and compared.

The mean differences for the target were -0.41 mm on the x axis (P < 0.001), 0.06 mm on the y axis (P = 0.412), and -0.34 mm on the z axis (P < 0.01). The mean absolute differences were 0.53 mm on the x axis (median, 0.50 mm; range, 0.00-2.00 mm), 0.46 mm on the y axis (median, 0.50 mm; range, 0.00-2.00 mm), and 0.78 mm on the z axis (median, 0.50 mm; range, 0.00-6.00 mm). The mean three-dimensional distance between MRI- and CT-derived coordinates for the anterior commissure was 1.65 mm, with a distance of more than 4 mm in two cases (2%). The mean three-dimensional distance for the posterior commissure was 1.65 mm, with a distance of more than 4 mm in two cases (2%). The mean three-dimensional distance for the target was 1.25 mm (median, 1.14 mm; range, 0.00-6.27 mm), with a distance of more than 4 mm in one case (1%).

Statistically significant but relatively small differences between MRI- and CT-derived target coordinates were found. In some cases (approximately 2% of this series), the differences between MRI- and CT-derived coordinates may be relatively large (greater than 4 mm and up to 8 mm). However, given the superior anatomic resolution of MRI and the nature of the stereotactic procedures under consideration, we conclude that MRI, when validated within an institution, may be used alone for target localization in pallidotomy and thalamotomy.