Program Information
Use of CyberKnife Logfiles to Aid the Derivation of PTV Margins for Liver Treatments Using Synchrony Respiratory Compensation
M Liu1*, E Vandervoort1,2 , J E Cygler1,2, (1) Department of Physics, Carleton University, Ottawa, ON, CA (2) The Ottawa Hospital Cancer Centre, Ottawa, ON, CA
Presentations
TH-CD-205-10 (Thursday, August 3, 2017) 10:00 AM - 12:00 PM Room: 205
Purpose: To assess the accuracy and estimate adequate PTV margins for CyberKnife® liver cancer treatments using the Synchrony® tracking system. Correlations between the linearity of Synchrony correlation models and the size of the treatment errors are also evaluated.
Methods: By analyzing treatment fraction log files generated by the Synchrony system for 72 liver patients, the total delivery error is calculated as the quadratic sum of three components: the correlation error (difference between the predicted and measured target position); the predictor error (due to a system latency) and the global beam positioning error (which can be measured using static phantoms). Possible margins are proposed which encompass the total delivery error for 90th, 95th, 99th percentile of all valid models, all patient fractions and the entire patient population. The relationships are also explored between total delivery error, amplitude and linearity of the internal to external breathing motion as well as with liver lobe in which internal markers are located.
Results: Considering either the 95th percentile of all models or all patients, 3.0 mm in superior-inferior direction, 1.7 mm in left-right direction, and 2.2 mm in anterior-posterior direction are sufficient to account for the total delivery error. Distinct clustering of models with low errors where internal-external motions are linearly related (95% have error < 3 mm, 80 % of the models have R squared > 0.72) indicates that Synchrony can predict well the target position for such cases. Weak correlation between peak-to-peak external and internal motions is visible but it cannot predict the size of errors as some poor models are interspersed with better models. The radial treatment errors across liver segments have considerable variation.
Conclusion: The treatment error spatial distribution suggests a non-isotropic PTV margin could be used for treatments and indicates that clinically-used 5 mm PTV margins might be overestimated.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by Ontario Consortium for Adaptive Interventions in Radiation Oncology (OCAIRO) grant.
Contact Email: