Room: AAPM ePoster Library
Purpose: To study the performance of our secondary monitor units (MU) calculation over time and evaluate the linear accelerator models and beam model changes over time.
Methods: 64,600 2ndary dose calculations over the last ten years of our independent verification software (RadCalc) have been accessed and categorized per linac model, treatment site and modality. From these 19,236 belong to modulated single beams or total dose of modulated beams. The linear accelerator beam models are those of Varian NovalisTX and ELEKTA VersaHD. The treatment planning system (TPS) used for all patient dose calculations was Pinnacle. The data was categorized in terms of treatment site, and beam models. Two beam models for each linear accelerator were created in the TPS over the period of calculations in an attempt to improve accuracy of dose calculations.
Results: On average, no difference was observed between linac models from the TPS. The average for the ELEKTA VersaHD models was 0.10±2.75%(initial) and 0.04%±2.64 (improved) and for the Varian NovalisTX were -0.81%±2.26 and -0.37%±2.08 respectively. Although a small improvement was observed in the averages before and after beam model improvements it was not statistically significant. Analysis of the 2nd MU dose calculations per anatomical site showed differences. Breast total dose calculations with RadCalc were -2.35±2.00% compared to the Pinnacle calculations and the rest of the sites were: abdomen 2.02±2.14%, pelvis 1.31±2.39%, brain 0.92±2.14%, thorax 0.96±2.52% and head-neck 0.30±2.37%.
Conclusion: Our historical data of secondary dose verification with RadCalc showed no statistically significant differences between initial TPS beam models and improvements of them. Although some differences are observed in the averages of dose calculations when grouped by treatment site, the differences are not statistically significant. Such differences can be explained by the differences in the dose calculation algorithms between TPS and RadCalc.