SU-E-J-259: Quantification of Rotational Localization Offset in Radiation Therapy




To verify the necessity of 6D localization detection and correction in radiation therapy.


An anthropomorphic head and neck phantom was used to test the BrainLab ExacTrac x-ray imaging system. After initial positioning, both ExacTrac and the on-board kV CBCT were used to detect the offset at the same position, using both manual and automatic registration algorithms. Then 6D offset including rotational errors up to 5 degree were introduced. Both ExacTrac and CBCT were used to correct the offsets and the residual errors were evaluated. Finally, 6D offset detected by ExacTrac for a C-Spine patient was reported.


The differences in 3D offset detected by ExacTrac and CBCT were 1.5 ± 1.2(Lateral), 2.7 ± 2.7(Vertical), and 4.0 ± 6.3(longitudinal) mm with manual registration while the corresponding differences decreased to 0.6 ± 0.3, 1.0 ± 0.3, and 0.3 ± 0.3 when automatic registration were used. CBCT corrected the translational offset to within 0.5 mm but the rotational errors remained and detected by the ExacTrac system (Yaw=2.1, Roll=1.1, Pitch=1.4 degree). When similar offset was introduced and corrected using ExacTrac, the residual error detected by both CBCT and ExacTrac were within 0.5 mm / 0.5 degree. The average offset from the 112 ExacTrac x-ray corrections for the C-Spine patient was 0.6 ± 1.6 (lateral), 5.4 ± 8(vertical), 1.6±1.1(longitudinal) mm, and 0.7 ± 0.6 (pitch), 0.7 ± 0.4(roll), 1.2 ± 0.7 (yaw) degree. Larger rotational errors, with a maximum of 2.7 degree (corresponds to about 1.5 to 4.5 mm offset for a POI 10 to 30 cm away from the isocenter), were observed when couch rotational were involved.


Rotational errors are common in patient localization, especially when couch rotation is involved. Both appropriate imaging system and 6D robotic couch are necessary to detect and correct the rotational localization errors.