Room: Exhibit Hall | Forum 4
Purpose: A software tool set to interrogate and harvest information from high resolution multi-leaf collimator (MLC) log files of the Elekta VersaHD. Secondly we will use the information to forecast MLC interventions and assess plan delivery quality.
Methods: A Matlab program was built in-house that performs comparisons for each logfile that has been generated by our VersaHD linear accelerator. The software reports the errors recorded during each beam delivery for the gantry location, collimator rotation, dose rate, jaws location and the location for each MLC. The minimum, maximum, median, average and standard deviation are calculated for each log file and each parameter. All parameters are also compared against the treatment planning system parameters and the differences are recorded. All calculated values are stored in a database for trending analysis.
Results: Our results show that there are minimal, average MLC errors after the analysis of 100,000 log files. Gantry angle errors range from -1 to 1 degree while negligible errors are reported for the collimator. Comparisons against the plan MLC locations show that most of MLC errors are observed in the periphery when the MLC leaves need to change direction.
Conclusion: Our software tool for the analysis of the Elekta log files has proven to be of great value because we can verify each beam delivery. Furthermore, the collected data can be used to identify highly-modulated plans and create guidelines to improve the plan optimization process to improve beam deliveries by minimizing the errors. Also, we will be using the data to monitor the performance of the linac and intervene when appropriate.