Program Information
Site Specific Analysis of Treatment Delivery Using MLC Log Files
J Rembish1*, D Defoor2 , N Papanikolaou3 , S Stathakis4 , (1) Texas A&M University, College Station, Texas, (2) ,,,(3) University of Texas HSC SA, San Antonio, TX, (4) Cancer Therapy and Research Center, San Antonio, TX
Presentations
TU-C1-GePD-TT-4 (Tuesday, August 1, 2017) 9:30 AM - 10:00 AM Room: Therapy ePoster Theater
Purpose: To quantify discrepancies in MLC positioning, gantry angle, and treatment delivery time during radiotherapy delivery of various treatment sites and linear accelerators.
Methods: The log files produced by the linear accelerator during treatment delivery were used to quantify delivery discrepancies. A database of the linac logfiles of all patients’ treatments was created using a homegrown Matlab software. MLC positions, gantry positions and treatment delivery time were extracted from over 100,000 MLC log files for over 600 patients treated using a Varian NovalisTx or an Elekta VersaHD. To date, only error in MLC positioning has been analyzed. Our software queries the database for various parameters. To determine the percent error, the planned position of the MLCs was compared to the actual position of the MLCs. The total error for each patient was divided by the total distance travelled by all MLCs.
Results: As expected, the amount of error increased as the intensity of the treatment plans increased. There was a linear correlation between MLC movement and MLC positioning error. Performing a Tukey simultaneous test with a 95% confidence interval revealed there was no significant difference in means of percent error, except between the chest and the brain. Data has only been analyzed for 56 patients so far and compared only to treatment site, however more comparisons are underway.
Conclusion: There does not appear to be any discrepancies in the amount of percent error that cannot be explained by the increased movement of the MLCs. The complexity of the plans as expressed in the total distance MLC travel is correlated to positional errors. Other variables, such as gantry positioning and exposure time are under investigation. In the future, our software solution should be able to provide ongoing statistics for the MLC performance and be used for preventive maintenance.
Contact Email: