The accuracy of magnetic resonance imaging for stereotactic localization

Stereotactic radiosurgery is based on the precise definition of intracranial target in relation to an extracranial reference system. A special plexiglas phantom for both magnetic resonance imaging (MRI) and computed tomography (CT) was built to test the spatial accuracy of MR images using the Leksell stereotactic head frame and localizer system. The phantom was constructed in a shape of cube, 164 mm in each dimension, with three perpendicular arrays of solid plaxiglas rod, 5 mm (3/16 inch) in diameter and spaced 30 mm apart within the phantom. In this study, both MRI and CT images were taken under the same Leksell head frame placement. The deviations of the intraphantom rod positions were determined by using fused image data sets from two MRI scanners and a CT scanner. For the Siemens MR scanner (Symphony, 1.5 T), mean differences of the coordinates in the rod position were 0.5 mm (range: 0.0 to 1.5 mm), 0.6 mm (range: 0.0 to 1.5 mm), and 0.3 mm (range: 0.0 to 1.0 mm) for the x, y and z coordinates of the Leksell stereotactic system, respectively. For the Philips MR scanner (Gyroscan, 1.5 T), mean differences were very similar to that of the Siemens. Least distortion appears in the middle of the phantom or the head frame system while most distortion appears around the periphery, MR images from these particular MR scanners have similar quality. The overall mean distortion is under 1 mm.