Background: In order to evaluate the possible submillimeter application accuracy in computer-aided navigation in the petrous bone, we performed a set of approximately 3,000 measurements on a specially prepared anatomic specimen using the Zeiss STN intraoperative navigation system. This allowed direct measurements of relevant anatomic structures in and around the petrous bone which are usually not directly accessible.
Results: We found that the best results can be achieved by exploiting contemporary multislice CT-imaging with 0.5 mm slice thickness and by direct radiologic imaging of the petrous bone; additionally, an extrinsic marker structure, the VBH-referencing element, served as an extension of the applied surface markers for the "patient-to-image" referencing procedure. Interestingly, the additional use of a surface registration, as provided by the STN-navigation system, to potentially optimize the registration, did not improve the results. In the best case, i.e. with high-resolution CT-imaging, 0.5 mm slice spacing, the use of surface markers, and the extrinsic referencing structure applied, an absolute difference between the calculated and actual position of the probe was 0.42+/-0.69 mm.
Conclusions: These results show that intraoperative 3-D navigation can be successfully transferred to a clinical application in the petrous bone or at the cerebellopontine angle with satisfactory accuracy in this highly sensitive anatomic region, even if only a restricted area of the patient can be used for the referencing procedure.