OBJECTIVE: To assess the accuracy of computer-aided surgery for common neurotologic approaches to the temporal bone. STUDY DESIGN AND SETTING: Cadaveric heads were dissected by using standard neurotologic approaches to the temporal bone including translabyrinthine, middle fossa, and retrosigmoid. Dissected anatomic landmarks from each approach were compared with CT images that were obtained before dissection on the VectorVision2 system (BrainLAB Corp, Munich, Germany) and the variability measured from digital images. Each approach was performed 5 times, with each anatomic landmark measured 3 times from each approach. RESULTS: The accuracy of the computer-aided surgery system was less than 1 mm for all anatomic points measured. Forty-two of the 49 measured points were accurate to less than 0.5 mm. CONCLUSIONS: Computer-aided surgery of the temporal bone using common neurotologic approaches can be performed accurately and reliably in cadaver specimens. SIGNIFICANCE: The utility and limitations of computer-aided surgery in the temporal bone are discussed. EBM rating: B-2.