gms | German Medical Science

Objective: High accuracy is crucial in stereotactic surgery as deviations >1,5 mm may be associated with poor outcome. Intraoperative fluoroscopy has been used to verify stereotactic targets and to survey implantation of deep brain stimulation (DBS) electrodes. In contrast to stereotactic x-ray, MRI and CT, there was a lack of phantom studies assessing accuracy of 3D fluoroscopy (XT) in stereotactic procedures. The objective was to assess the accuracy of XT in phantom testing and in a clinical series compared to computed tomography (CT) as gold standard.

Method: The sphere phantom consists of 3x5 metallic spheres with defined distances. Thus 105 known distances were available to verify with XT and CT. In the skull phantom 3 drilled trajectories contained tubes with 13 metallic spheres in defined distances. XT and CT were performed with and without a mounted stereotactic head-frame, placing the phantom with different angulations into the gantry to simulate variable patient positioning. We also analysed 8 patients (16 trajectories) after DBS surgery, assessed by intraoperative XT and postoperative CT. Primary endpoint was accuracy of distance measurements on XT compared to CT, assessing means, mean squared error and standard deviation. Significance concerning accuracy was tested with Friedman's test, significance concerning precision was tested with Levene's test. Secondary endpoints were target accuracy by calculating the Euclidian error and angles demonstrated by CT and XT in the skull phantom and the patient cohort.

Results: The mean error between known and measured distance in the sphere phantom using XT scan was 0,33 mm, compared to 0,19 mm for CT scans, p=0.0004. The differences in standard deviation and dispersion were significant (p<0.0001). In the skull phantom the mean error of XT was 0,28 mm, compared to 0,39 mm in CT, p=0.59. Differences in deviation and dispersion were significant in Levene's test, p<0.0001. A total of 15 trajectories were measured with CT and XT. Mean difference was 0,3 mm (x–axis), 0,4 mm (y–axis), 0,43 mm (z–axis), 0,09 mm (target length), 0,91° (ring angle) and 0,58° (arc angle). In the DBS cohort 16 trajectories were evaluated. Mean difference was 0,51 mm (x–axis), 0,58 mm (y–axis), 0,76 mm (z–axis), 0,72° (ring angle) and 0,61° (arc angle).

Conclusions: XT showed a lower accuracy but higher precision than CT. The accuracy of XT is however higher than the application accuracy of stereotactic frames. Therefore XT should be acceptable in stereotactic settings.