Purpose: To develop an accurate method to incorporate rotational setup differences for QACT dosimetric evaluations using Eclipse and Precision treatment planning systems.
Methods: Recalculation of the original treatment plan on a CBCT or updated fan-beam CT (QACT) is often used to provide the dosimetric impact of anatomical changes and to decide whether or not an adaptive plan is necessary. However, some existing technologies do not account for rotational differences (roll, pitch, and yaw) between the QACT and the planning CT. The VelocityAI software was used to resample the QACT to the planning CT frame-of-reference, to simulate rotational corrections that are applied at the treatment machines (up to 3° for TrueBeam and 5° for CyberKnife). Effect of omission of rotations on QACT dose estimates was studied on three pelvic treatment cases for TrueBeam/Eclipse with para-aortic and inguinal nodal targets and on an intracranial lesion treatment for CyberKnife/Precision. The original treatment plans were recalculated on the QACTs with and without rotational differences, to demonstrate the dosimetric sensitivity.
Results: All three pelvic cases showed clinically unacceptable (<91% of prescription dose) minimum D0.03cc doses when rotational differences were not accounted for (63.5%-88.0%). Target coverage was acceptable when rotations were applied (minimum D0.03cc = 91.8%-95.8%). The CyberKnife case showed a small increase in optic nerve doses and a 16% increase in mean right cochlea dose when rotational differences were not taken into account. Dosimetric differences for OARs were negligible when rotations were applied.
Conclusion: All four clinical examples presented would have triggered adaptive replanning if rotational differences in the patient CTs were not taken into account. Application of rotational corrections applied at the treatment machines is an important parameter when assessing the dosimetric impact of anatomical changes, and is a crucial component to adaptive RT workflows.