Dose calculation errors for volumetric modulated arc therapy plans of various complexity

BACKGROUND AND PURPOSE: Volumetric modulated arc therapy (VMAT) plans often involve small and irregular beam apertures (i.e. high complexity), which can pose challenges for accurate dose calculations. The aim of this study was to investigate plausible dose calculation errors for VMAT plans of various complexity. Patient/material and methods: Twenty patient cases, each with three VMAT plans of different complexities (i.e. 60 plans in total), were included. All plans were calculated using six different dose calculation methods. It was assumed that greater differences between calculations reflect increased difficulty in accurately estimating dose. Basic performances of the six calculation methods were evaluated based on static fields. Three-dimensional distributions of voxel-wise two standard deviations (2SDs) in percent of local voxel dose were visually evaluated. 2SD volume histograms were analyzed as well as the mean (2SDMean) and maximum 2SD within 2 cm3 (2SD2cc) for different regions of interest. RESULTS: Higher 2SD values were generally found outside the planning target volume (PTV) compared to inside, particularly in low dose and buildup regions. Average (per treatment site) 2SDMean/2SD2cc values were 0.9-1.2%/1.8-5.1% in the PTV and 1.8-3.1/8.4-18.7% for the region 1 cm outside the PTV. For organs of interest, 2SDMean/2SD2cc values were larger than the equivalent values in the PTV, with highest values up to 5/15% observed for the prostate cases. INTERPRETATION: Variations between dose calculation methods were larger in organs of interest than in the PTVs. Differences in 2SD distributions between the patient cases were generally larger than differences between the complexity levels.