Room: AAPM ePoster Library
Purpose: A general analytical tool is introduced to evaluate the dosimetric impacts of delivery errors for CyberKnife radiosurgery treatments. Two very different types of treatments are evaluated, liver cancer and trigeminal neuralgia (TN), to demonstrate its broad applicability.
Methods: The 3D contours, planned dose and treatment delivery log-files are extracted for 148 liver and, as a proof of principle, for 13 left-sided TN treatments. For liver targets using Synchrony tracking, the errors considered are translational motion-compensation errors, rotations which were too large for the system to correct and target deformations. The dosimetric parameter evaluated for liver is dose to 99% of the GTV volume (D99) divided by the prescribed dose (Rx). For skull tracking based TN treatments, translational and rotational errors are evaluated assuming rigid body geometry. The dosimetric parameters considered are length (L) of trigeminal nerve receiving the prescription dose, and dose to 0.03 cc of the brainstem. The dosimetric sensitivity of plans to uncorrected rotations is tested for both anatomical sites.
Results: For liver treatments, only 47% of patients had rotations tracked by the system, mainly due to imperfect fiducial placement. After accounting for delivery error in patients with rotations, D99/Rx is < 100% for 33% of targets. For all patients, the 25th percentile of D99/Rx is > 98% for uncorrected rotations below 3 degrees. The distributions of D99/Rx are broad even for the planned dose indicating that coverage is highly patient dependent and often compromised for organs at risk. For TN treatments taking into account delivery error, L > 4 mm and V15Gy for brainstem is < 0.03 cc for all patients. The dosimetric impact of rotations for TN is highly dependent on the sign and direction of the uncorrected rotation.
Conclusion: The dosimetric tool described can evaluate margin suitability and rotation sensitivity for radiosurgery plans.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by a research grant from Accuray, Inc., the Ontario Consortium for Adaptive Interventions in Radiation Oncology [File # RE-04-026], an Ontario Graduate Scholarship and the John Lyndhurst Kingston Scholarship.