Evaluation of Patient Specific Machine Delivery Performance Based On Analysis of Trajectory Log Files
V Stakhursky1*, T Stanley2, W Yi2, (1) Steward Health Care, Methuen, MA, (2) CCS Oncology, Buffalo, NYSU-E-T-380 Sunday 3:00:00 PM - 6:00:00 PM Room: Exhibit Hall
Purpose: Comparison of delivered and treatment planning dose to a phantom is routinely done before first patient treatment for IMRT and VMAT treatment courses, and is an important part of QA program. It is assumed that the delivery parameters are reproducible for each treatment in the course, and additional tests to validate the assumption are rarely performed. We use the analysis of Trajectory log files to verify the interfractional constancy of treatment delivery.
Methods: Trajectory Log files present a record of machine coordinate snapshots every 10ms of beam on state. We designed an automated data mining engine capable of extracting machine parameters recorded for the treated fractions for a selected patient, and generating easy to review statistical report quantifying machine delivery performance. The application was tested on 3 head and neck and 4 pelvis test courses in our clinic.
Results: The worst observed instant error in Gantry angle position during the delivery of VMAT treatments was 0.22deg, while the time-averaged gantry error during the delivery was less than 0.002deg. The maximum observed discrepancy between planned and actual treatment position for MLC leaves was 0.11mm. When machine parameter interfractional variance was analyzed, the worst difference in gantry position between any fraction delivered for a patient was 0.013deg. The worst instantaneous discrepancy between any fraction of treatment for MLC leaves was only 0.028mm.
Conclusions: The analysis of Varian TrueBeam accelerator Trajectory Log files generated during the course of treatment of 3 head and neck and 4 pelvis patients showed high reproducibility in dynamics of machine position parameter during the delivery of treatments. Continuous monitoring of treatment logs enables verification of treatment delivery constancy, and should be integrated in clinical QA programs.