Room: AAPM ePoster Library
Purpose: Monte Carlo (MC) simulation directly calculates the dose as dose = (energy deposited) / (physical mass), while AAA uses convolution/superposition based on energy kernels. These kernels are re-scaled using equivalent electron-density path for heterogeneous geometry. This study investigates the difference of dose reported by the AAA algorithm for high density materials compared to those calculated by MC.
Methods: Spine SBRT patients were used. VMAT plans were generated in Varian’s Eclipse. The same plan was converted to a MC model in our commissioned EGS4-based MC system, MCSIM. Region-of-interest (ROI) structures (0.2cm³ each) were created at different iso-dose-line (IDL), identified as ROI_LD (~ 50%Rx IDL), ROI_HD1 (~126%Rx IDL), and ROI_HD2 (~116%Rx IDL). Dose comparisons between MC and AAA (ver.13.6, Varian) were performed under the following scenarios: (1) the whole patient body was assigned as water; (2) the whole body was assigned as water except ROI_HD1 as medium-density bone (density =1.83 g/cm³, relative electron density = 1.71); (3) the whole body was assigned as water except ROI_HD1 as titanium (Ti) (density =4.51 g/cm³, relative electron density = 3.72). For MC, a 1mm voxel size was used with ICRU cross-section data and ECUT = 0.7 MeV, PCUT = 0.01 MV, and a 1-s dose uncertainty <1%. A dose ratio DR=(AAA dose)/(MC dose) was calculated.
Results: For scenario 1, DVHs match perfectly with mean DR =100.1% for ROI_LD, 100% for ROI_HD1, and 100% for ROI_HD2. For scenario 2, these DR values are the same for ROI_LD and ROI_HD2, except for ROI_HD1 (DR=101.6%). For scenario 3, these DR values change little for ROI_LD (DR =100.2%) and ROI_HD2 (DR =99.8%), except for ROI_HD1 (DR =105.6%).
Conclusions: The results show that the reported dose by AAA for high-density media is much higher than MC calculation (e.g., 1.6% higher in medium-density bone and 5.6% in titanium).
Dose, Monte Carlo, Treatment Planning