SU‐E‐T‐34: A Feasible Action Level System for DVH‐Based IMRT QA

Purpose:Gamma‐based IMRT QA metrics have been shown to inadequately predict clinically relevant patient dose errors. In this study, we develop and evaluate an intuitive and feasible DVH‐based IMRT QA action level system, which incorporates both physical and clinical considerations.Methods:VMAT plans from previously treated prostate (n=11) and brain (n=13) patients are analyzed. Treatment plans were created using Eclipse TPS and delivered using Varian's Trilogy linac. Initial IMRT QA was performed using the ArcCHECK phantom. All plans passed gamma‐based criteria (3%/3mm) with at least a 97% passing rate. To design and test the feasibility of the DVH‐based IMRT QA metric, 3DVH software is used to process the QA data from the ArcCHECK measurement and predict the 3D distribution of dose delivered to the patient. This enables comparison of a standard set of clinically important DVH values between TPS calculations and that predicted by the QA process. A two‐level system is implemented to determine the IMRT QA Result. Level 1 requires that QA predicted DVH values for the PTV and OARs fall within a physical tolerance generated by scaling the original planned dose uniformly by ±5%. If level 1 tolerance is exceeded, level 2 requires that the QA‐predicted DVH values are within institutional normal tissue tolerances. OARs for prostate cases include bladder and rectum. OARs for brain cases include the lens, optic nerve, optic chiasm, and brainstem.Results:All analyzed QAs passed the proposed action levels, with prostate cases all satisfying physical tolerances. Ten brain cases had DVH values failing physical tolerances, but satisfying clinical tolerances. All DVHs failing physical tolerances are in low dose region.Conclusion:The proposed IMRT QA metrics represents a straightforward alternative to the gamma‐based metric. It has the potential to better address both physical and clinical concerns that are specific to individual patients.