Robust Prediction of Clinical Deep Brain Stimulation Target Structures via the Estimation of Influential High-Field MR A
This work introduces a robust framework for predicting Deep Brain Stimulation (DBS) target structures which are not identifiable on standard clinical MRI. While recent high-field MR imaging allows clear visualization of DBS target structures, such high-fi
- PDF / 787,875 Bytes
- 8 Pages / 439.363 x 666.131 pts Page_size
- 41 Downloads / 166 Views
Department of ECE, Duke University, Durham, NC, USA CMRR, University of Minnesota, Minneapolis, MN, USA 3 Department of Biomedical Engineering, Duke University, Durham, NC, USA 4 Department of Neurology, University of Minnesota, Minneapolis, MN, USA 5 Department of Neurosurgery, University of Minnesota, Minneapolis, MN, USA 2
Abstract. This work introduces a robust framework for predicting Deep Brain Stimulation (DBS) target structures which are not identifiable on standard clinical MRI. While recent high-field MR imaging allows clear visualization of DBS target structures, such high-fields are not clinically available, and therefore DBS targeting needs to be performed on the standard clinical low contrast data. We first learn via regression models the shape relationships between DBS targets and their potential predictors from high-field (7 Tesla) MR training sets. A bagging procedure is utilized in the regression model, reducing the variability of learned dependencies. Then, given manually or automatically detected predictors on the clinical patient data, the target structure is predicted using the learned high quality information. Moreover, we derive a robust way to properly weight different training subsets, yielding higher accuracy when using an ensemble of predictions. The subthalamic nucleus (STN), the most common DBS target for Parkinson’s disease, is used to exemplify within our framework. Experimental validation from Parkinson’s patients shows that the proposed approach enables reliable prediction of the STN from the clinical 1.5T MR data.
1
Introduction
Deep brain stimulation (DBS) surgery is commonly used for symptom’s treatment in neuro-degenerative diseases such as Parkinson’s disease (PD). Precise placement of electrodes within crucial sub-cortical region (e.g., subthalamic nucleus (STN)) leads to successful DBS procedures [1]. Standard DBS targeting approaches today refer to anatomical information based on normalized atlases, particularly based on a single histology sample [1–3]. However, in such indirect methods, the variability in the position and size of the DBS targets needs to be further analyzed in the context of large populations for the reliability [2, 3]. To verify the target location, electrophysiological measurements, such as microelectrode recording (MER), that are lengthy and might result in increased risks for hemorrhage are required during surgery [1–3]. c Springer International Publishing Switzerland 2015 N. Navab et al. (Eds.): MICCAI 2015, Part II, LNCS 9350, pp. 587–594, 2015. DOI: 10.1007/978-3-319-24571-3_70
588
J. Kim et al.
Direct visualization and localization of the targets on the individual patient are needed for reliable, safe, and time-efficient DBS targeting. With advances in high-field MR (e.g., 7 Tesla (7T)), the superior contrast and high resolution imaging allow to directly identify and visualize the DBS targets, reducing the need for MER and other lengthy intra-operative burdensome steps [1, 3–6]. However, such high-fields are limited in clinical use, and thus the ta
Data Loading...
