CARS 09 Selbekk abs

From IGSTK

Freehand 3D ultrasound and navigation in neurosurgery - technical requirements

T. Selbekk1,2, 4, F. Lindseth 1,2, I Reinertsen 1,2, 4, O. Solheim2,3,4, G. A. Tangen1,2, O. V. Solberg1,2,4, L. E. Bø1,2, Ø. Nygård3,4, TAN Hernes1,2, 4, G. Unsgård2,3,4 ,

1SINTEF, Medical Technology, Trondheim, Norway
2National centre for 3D ultrasound in surgery, Trondheim, Norway
3St. Olav University Hospital, Neurosurgery, Trondheim, Norway
4Norwegian University of Science and Technology, Trondheim, Norway

Keywords: ultrasound, neurosurgery, navigation

Purpose The purpose is to describe the technical setup and requirements of intraoperative ultrasound and navigation technology for various neurosurgical procedures, including surgery of cerebral tumours and cerebral aneurysms or arterio-venous-malformations (AVM), spinal tumour surgery and transphenoidal surgery of pituitary tumours. Clinical feasibility of the in-house research navigation system CustusX and high-resolution ultrasound will be demonstrated.

Method The technical system in the operating room (OR) consists of a ceiling-mounted optical tracking camera (NDI, Waterloo, Canada), and the in-house research navigation software CustusX installed on a Macbook computer (Apple, Cupertino, USA). The ultrasound images from a commercial ultrasound scanner were transferred to the navigation system through a standard S-VHS cable, reconstructed to a 3D volume and displayed, enabling navigation based on the intraoperative ultrasound volumes. Different ultrasound probes were used, depending on the clinical procedure. In spine surgery a commercial multi-row flat linear array 8-14 MHz was used, while in pituitary surgery a prototype probe manufactured according to our specifications was used.

Results The 2D ultrasound provided high-resolution images, often with image quality superior to the preoperative images. In pituitary surgery, the ultrasound image size was typically 13x20 mm, while in spine surgery it was approx. 45x50 mm. Typically 200-300 ultrasound images were assembled in an image volume, and voxel sizes down to 0.1 mm were achievable in the 3D reconstruction. The ultrasound volumes covered the area of interest only, which in pituitary surgery means a volume size of approx. 13 x 20 x 20 mm. Conclusion Navigation based on high-resolution 3D ultrasound image volumes was feasible using the described setup. The high-resolution ultrasound images and the small size of the image volume put higher demands on the probe calibration, the 3D reconstruction and the visualization of image volumes than in conventional ultrasound based navigation.