Program Information
Validation of 3DVH Accuracy On Quantifying Delivery Errors Based On Clinical Relevant DVH Metrics
T Ma*, L Kumaraswamy , Roswell Park Cancer Institute, Buffalo, NY
Presentations
SU-E-T-789 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: Detection of treatment delivery errors is important in radiation therapy. However, accurate quantification of delivery errors is also of great importance. This study aims to evaluate the 3DVH software’s ability to accurately quantify delivery errors.
Methods: Three VMAT plans (prostate, H&N and brain) were randomly chosen for this study. First, we evaluated whether delivery errors could be detected by gamma evaluation. Conventional per-beam IMRT QA was performed with the ArcCHECK diode detector for the original plans and for the following modified plans: (1) induced dose difference error up to ±4.0% and (2) control point (CP) deletion (3 to 10 CPs were deleted) (3) gantry angle shift error (3 degree uniformly shift). 2D and 3D gamma evaluation were performed for all plans through SNC Patient and 3DVH, respectively. Subsequently, we investigated the accuracy of 3DVH analysis for all cases. This part evaluated, using the Eclipse TPS plans as standard, whether 3DVH accurately can model the changes in clinically relevant metrics caused by the delivery errors.
Results: 2D evaluation seemed to be more sensitive to delivery errors. The average differences between ECLIPSE predicted and 3DVH results for each pair of specific DVH constraints were within 2% for all three types of error-induced treatment plans, illustrating the fact that 3DVH is fairly accurate in quantifying the delivery errors. Another interesting observation was that even though the gamma pass rates for the error plans are high, the DVHs showed significant differences between original plan and error-induced plans in both Eclipse and 3DVH analysis.
Conclusion: The 3DVH software is shown to accurately quantify the error in delivered dose based on clinically relevant DVH metrics, where a conventional gamma based pre-treatment QA might not necessarily detect.
Contact Email: