Borders of STN determined by MRI versus the electrophysiological STN. A comparison using intraoperative CT

Background It is unclear which magnetic resonance imaging (MRI) sequence most accurately corresponds with the electrophysiological subthalamic nucleus (STN) obtained during microelectrode recording (MER, MER-STN). CT/MRI fusion allows for comparison between MER-STN and the STN visualized on preoperative MRI (MRI-STN). Objective To compare dorsal and ventral STN borders as seen on 3-Tesla T2-weighted (T2) and susceptibility weighted images (SWI) with electrophysiological STN borders in deep brain stimulation (DBS) for Parkinson's disease (PD). Methods Intraoperative CT (iCT) was performed after each MER track. iCT images were merged with preoperative images using planning software. Dorsal and ventral borders of each track were determined and compared to MRI-STN borders. Differences between borders were calculated. Results A total of 125 tracks were evaluated in 45 patients. MER-STN started and ended more dorsally than respective dorsal and ventral MRI-STN borders. For dorsal borders, differences were 1.9 +/- 1.4 mm (T2) and 2.5 +/- 1.8 mm (SWI). For ventral borders, differences were 1.9 +/- 1.6 mm (T2) and 2.1 +/- 1.8 mm (SWI). Conclusions Discrepancies were found comparing borders on T2 and SWI to the electrophysiological STN. The largest border differences were found using SWI. Border differences were considerably larger than errors associated with iCT and fusion techniques. A cautious approach should be taken when relying solely on MR imaging for delineation of both clinically relevant STN borders.