Clinical use of a frameless stereotactic arm: results of 325 cases.

The viewing wand is a frameless stereotactic arm that can be used in conjunction with computerized tomography (CT) or magnetic resonance (MR) imaging to provide image-based intraoperative navigation. The authors report a series of 325 cases in which the viewing wand was used and evaluated for its utility, ease of integration into the standard surgical setup, reliability, and real-world accuracy. The use of the system was associated with minimal additional effort or time spent in setting up the procedure as long as a trained technician performed the data transfer and reconstruction. The viewing wand was used in 165 cases in conjunction with CT and 145 cases with MR imaging. The system was reliable, achieving a useful registration in 310 of 325 cases (95.4%). Fiducial-based registration was more accurate than an anatomical landmark-surface fit algorithm method of registration (mean 2.8 vs. 5.6 mm error, respectively, for CT; and mean 3.0 vs. 6.2 mm for MR imaging). The actual error of the system in estimating the position of the probe tip just after registration was judged by the operating surgeon to be less than 2 mm in 92% of MR imaging cases and in 82% of CT cases, between 2 and 5 mm in 7% of MR imaging and 17% of CT cases, and greater than 5 mm in less than 1% of MR imaging and 1.2% of CT cases. The accuracy of the system degraded during the operation, so that by the third evaluation the error was estimated to be less than 2 mm in 77% of MR imaging and 62% of CT cases. Overall, the viewing wand was found to be reliable and accurate. This real-world accuracy was sufficient for a broad range of applications including glioma resection, cerebrospinal fluid shunting procedures, resection of small subcortical masses, and temporal lobe resection. The system is a useful navigational aid that allows a direct approach to intracranial pathology without the drawbacks of application and the limitations of a stereotactic frame.