Abstract

In neurosurgery, navigation is being used to improve surgical orientation by using preoperative images as a roadmap. Skin or bone fiducials couple the image coordinate system to that of the patient’s head fixed by the Mayfield clamp. Then the tip of a pointer of another instrument (localization device) can be seen in relation to the image to give the surgeon insight where he/she is in the brain and where the tumor or lesion can be expected in the depth.

Drawbacks from current navigation systems are that 1) they only show the actual position of the localization device and thus do not hint whether the surgeon has removed the tumor completely, 2) don’t warn when the device is about to hit a critical brain structure, and 3) do not compensate for shifts of the brain during surgery invalidating the pre-operative image data.

During the last 5 years we investigated in our hospital whether sound and workflow feedback could improve the surgical resection accuracy and looked how the pre-operative image data could be deformed in real-time using GPU hardware to match the tracked cortical surface to compensate for brain shifts.

© 2011 OSA/SPIE