Video-based Navigation for Endoscopic Sinus Surgery
It is estimated that there are more than 200,000 functional endoscopic sinus surgeries (FESS) procedures performed annually in the United States at a cost of several billion dollars annually, all of which are performed under endoscopic guidance, and a large fraction of which employ surgical navigation. However, the accuracy of current navigation technology is 2 mm under near ideal conditions, and as a result, confirmation of sensitive anatomic structures typically relies on the surgeon’s ability to interpret and relate CT images to the endoscopic view. The goal of this project is to overcome these limitations by improving registration accuracy with the viewing endoscope through in-situ registration of CT imagery to intraoperative endoscopic imagery. This then allows computation of the shape of viewed surfaces directly from endoscopic video sequences with the goal of improved abilities to assess the progress of surgery and a reduced need for intraoperative imagery. Together, these capabilities will provide enriched visualizations and enhance patient safety, reduce complications, and improve clinical workflow.
We have performed experiments in simulated data as well as in-vivo patient data from both the OR and clinic. We are currently preparing our software for deployment into a clinical trial.