Virtual reality systems can be used for the training of surgeons. They are extremely beneficial for situations where endoscopic surgery is the method of choice. This is the case in diagnostics as well as in the operative therapy of diseases of the nasal and paranasal cavities and their anatomically adjoining regions. There the endoscope renders valuable services.

The eye of the surgeon is directly brought to the pathological structure. By means of angular optics, the inspection of regions is possible which are otherwise only accessible by generous exploration. The reduced invasiveness is however attained at the expense of a limited overview of the field of operation. Orientation is therefore accomplished by means of certain anatomical landmarks which generally provide a sufficient determination of the instrument's position in the situs. In the case of revision operations with lacking landmarks and intense cicatrization, orientation can be difficult even for experienced surgeons. A higher risk of damaging orbital, neural and vascular structures also exists for anatomical variants. In these cases even the routinely prepared computertomographic scans cannot guarantee intra-operative orientation.

Functional nasal surgery includes septoplasty and rhinoseptoplasty, interventions that are carried out to improve nasal airflow. One out of two patients that undergo this treatment is dissatisfied with the outcome. Surprisingly patient satisfaction is not correlated with mere post-operative nasal patency nor with measurements of nasal airflow carried out under forced ex- and inspiration. Based on a 3:1 enlarged physical model that was hand-made from a CT volume model of one patient's nasal cavity we used the laser photosection technique, which can be analysed plane by plane with particle image velocimetry (PIV) to determine the parameters of flow within the nose. These parameters are input to a numerical model for the nasal airflow based on finite element analysis. Results of this analysis comply very well with measurements of nasal airflow using rhinomanometry. We can demonstrate, that during physiologic breathing conditions, turbulence and even recirculation may occur in noses with in general sufficient cross sectional patency. This may induce the sensation of nasal blockage in some patients following nasal surgery. Also serious complications as are reported in literature may occur.

In a next step we have designed a surgeon's workbench realised on a CAVE-device, where the surgeon can interactively modify the CT-based patient model. The result of this virtual endoscopic intervention is the input for the numerical model. This allows for iterative optimisation of individual anatomy. In my talk I am going to evaluate the possibilities of combining this model with intraoperative position measurement in endonasal endoscopic surgery.