gms | German Medical Science

Objective: Fiducial-based registration for navigation has an overall clinical application accuracy of 3-4 mm. The aim of this study is to increase navigation accuracy by being less user-dependent applying an automatic registration approach by intraoperative computed tomography (CT).

Methods: In a series of 26 patients (m/f: 12/14, age range: 53.65±16.96 years) operated on intracranial lesions we applied a 32 slice CT scanner (AIRO, Brainlab, Munich) for pre- and intraoperative imaging, automatic navigation registration and microscope based navigation. Reflective markers on the CT scanner in combination with a detachable CT-compatible head reference star allowed automatic navigation registration. A rigid registration algorithm was further applied to register the CT data with preoperative navigation planning, based on 3T MRI. Navigation accuracy was measured by three independent skin fiducials not used for the registration process and additionally four acquired bony landmarks around the craniotomy opening. Initial registration accuracy and re-registration accuracy was determined. For pre- and intraoperative imaging after tumor resection the reference star is exchanged for an unsterile reference star, the sterile one is re-attached after scanning, so additionally registration accuracy of the reference star itself was measured.

Results: Navigation accuracy prior to skin incision documented by skin fiducials proved to be less than 2 mm in all cases. The registration accuracy of the re-mounted reference star varied for initial registration from 0.6 to 2.1mm (mean±sd: 1.48±0.61mm) and for re-registration from 0.5 to 2.0mm (mean±sd: 1.27±0.62mm). Bony landmark accuracy after intraoperative imaging ranged from 0.1 to 1.4mm (mean±sd: 0.58±0.37mm) for initial registration and 0.2 to 1.9mm (mean±sd: 0.94±0.47m) for re-registration.

Conclusion: Intraoperative CT-based automatic registration increases navigation accuracy compared to the standard fiducial based user-dependent approach.