Geometric accuracy in patient positioning for stereotactic radiotherapy of intracranial tumors.

Background/Purpose: This study determines and compares the geometric setup errors between stereoscopic x-ray and kilo-voltage cone beam CT (CBCT) in phantom tests on a linear accelerator (linac) for image-guided (IG) stereotactic radiotherapy of intracranial tumors. Additionally, dose-volume metrics in the target volumes of the setup errors of CBCT were evaluated. Materials/Methods: A Winston-Lutz- and an anthropomorphic phantom were used. The mean deviation and root mean square error (RMSE) of CBCT and stereoscopic x-ray were compared. Dose-volume metrics of the planning target volume (PTV) and gross target volume (GTV) for CBCT were calculated. Results: The RMSEs in the tests with the Winston-Lutz-Phantom were 0.3 mm, 1.1 mm and 0.3 mm for CBCT and 0.1 mm, 0,1 mm and <0.1 mm for stereoscopic x-ray in the translational dimensions (right-left, anterior-posterior and superior-inferior). The RMSEs in the tests with the anthropomorphic phantom were 0.3 mm, 0.2 mm and 0.1 mm for CBCT and 0.1 mm, 0,1 mm and <0.1 mm for stereoscopic x-ray. The effects on dose-volume metrics of the setup errors of CBCT on the GTV were within 1 % for all considered dose values. The effects on the PTV were within 5 % for all considered dose values. Conclusion: Both IG systems provide high accuracy patient positioning within a submillimeter range. The phantom tests exposed a slightly higher accuracy of stereoscopic x-ray than CBCT. The comparison with other studies with a similar purpose emphasizes the importance of individual IG installation quality assurance.