SU-E-T-690: Sensitivity Analysis to Machine Error: Comparisons of VMAT QA Systems.

PURPOSE: In VMAT delivery, the gantry angles, dose rate and the MLC positions are all modulated. It is therefore necessary to perform pre-treatment QA to detect any possible errors either in the plan or machine delivery system. Previously we developed a QA technique for VMAT using EPID in cine mode on a TrueBeam LINAC. The purpose of this study is to compare the technique with commercially available VMAT QA systems on the sensitivity to machine error.

METHODS: In our technique, the VMAT plan was delivered with EPID acquiring cine images simultaneously. The measured portal dose (PD) is derived from the acquired images. The predicted PD is calculated using the control point information in the plan and the delivery information in the EPID images. 3D γ analysis with the 3rd axis being gantry angle is performed between predicted and measured PD. The commercial VMAT QA system was Delta-4 system from ScandiDose. The QA of the original plan is the reference. Several types of machine errors were simulated and evaluated: gantry angle, MLC position, and output. The original plan is modified by introducing intentional errors of various magnitudes. The modified plan is delivered and the measured PD is compared with the original predicted PD to determine the sensitivity to machine errors.

RESULTS: For the reference, in Delta-4, the γ passrate is 99.6% for 3%/3mm criteria, and our technique has γ passrate of 98.5% for 3%/3mm/2° criteria. When errors are introduced, the passrate decreases. The sensitivity of Deltat4 (EPID) to gantry angle, MLC, and output are 1.4%/° (2.8%/°), 22%/mm(6%/mm), and 0%/% (0.1%/%), respectively. Similar results were observed under different criteria.

CONCLUSION: Compared with Delta-4 system, the EPID technique is more sensitive to gantry angle error, but less sensitive to MLC positioning error. Delta-4 system is insensitive to the machine output error. This work was partially supported by the National Natural Science Foundation of China (61171005) and the China Scholarship Council (CSC).