Error analysis of MRI and leksell stereotactic frame target localization in deep brain stimulation surgery

Stereotactic deep brain stimulation (DBS) is the surgical treatment of choice for medication-refractory patients with Parkinson's disease and essential tremor. The subthalamic nucleus and ventral intermediate nucleus of the thalamus appear to be effective targets for electrode placement. Because these targets are small and encased in fiber tracts, their localization can be exceedingly difficult. However, the precision of electrode placement is crucial for obtaining successful results. Currently, surgeons rely on preoperative MRI or CT images to derive stereotactic coordinates for targeting sites such as the subthalamic nucleus and ventral intermediate nucleus of the thalamus coupled with microelectrode recordings during surgery for proper electrode placement. However, it has been argued that the stereotactic head frame produces detrimental artifacts during MRI. We examined MRI images taken from 11 patients undergoing repeat DBS surgery, and determined the coordinates of the previously placed electrode. We then set the Leksell G stereotactic frame to these coordinates and obtained fluoroscope-localizing images. Using MATLAB image analysis tools, we were able to quantify the 3-dimensional error in target localization by measuring the distance from the electrode tip to the targeted coordinate. The mean errors were 0.09 +/- 0.34 mm perpendicular, lateral to medial, 0.01 +/- 0.32 perpendicular, posterior to anterior, and -0.08 +/- 0.33 mm parallel to the electrode, superior to inferior. According to statistical analysis, the error was random and did not seem to move in any predictable fashion. Therefore, we conclude that preoperative MRI images can be safely used in DBS surgery, and they do not negatively affect its accuracy.