Accurate CT-MR image registration for deep brain stimulation: a multi-observer evaluation study

ABSTRACT Since the rst clinical interventions in the late 1980s, Deep Brain Stimulation (DBS) of the subthalamic nucleushas evolved into a very e ective treatment option for patients with severe Parkinson's disease. DBS entailsthe implantation of an electrode that performs high frequency stimulations to a target area deep inside thebrain. A very accurate placement of the electrode is a prerequisite for positive therapy outcome. The assessmentof the intervention result is of central importance in DBS treatment and involves the registration of pre- andpostinterventional scans.In this paper, we present an image processing pipeline for highly accurate registration of postoperative CTto preoperative MR. Our method consists of two steps: a fully automatic pre-alignment using a detection of theskull tip in the CT based on fuzzy connectedness, and an intensity-based rigid registration. The registrationuses the Normalized Gradient Fields distance measure in a multilevel Gauss-Newton optimization frameworkand focuses on a region around the subthalamic nucleus in the MR.The accuracy of our method was extensively evaluated on 20 DBS datasets from clinical routine and comparedwith manual expert registrations. For each dataset, three independent registrations were available, thus allowingto relate algorithmic with expert performance. Our method achieved an average registration error of 0 :95mmin the target region around the subthalamic nucleus as compared to an inter-observer variability of 1 :12mm.Together with the short registration time of about ve seconds on average, our method forms a very attractivepackage that can be considered ready for clinical use.Keywords: deep brain stimulation, image registration, normalized gradient elds, multimodal registration,image fusion, CT, MRI