Assessment of Eclipse electron Monte Carlo output prediction for various topologies

Abstract

Monte Carlo simulation is deemed to be the leading algorithm for accurate dose calculation with electron beams. Patient anatomy (contours and tissue densities) as well as irradiation geometry is accounted for. The accuracy of the Monitor Unit (MU) determination is one essential aspect of a treatment planning system. Patient-specific quality assurance of a Monte Carlo plan usually involves verification of the MUs with an independent simpler calculation approach, in which flat geometry is to be assumed. The magnitude of the discrepancies between flat and varied surfaces for a few scenarios has been investigated in this study. The ability to predict MUs for various surface topologies by the commercial electron Monte Carlo implementation from Varian Eclipse system (Eclipse eMC) has been evaluated and compared to the Generalized Gaussian Pencil Beam (GGPB) algorithm. Ten phantoms with different topologies were constructed of water-equivalent material. Measurements with a parallel plate ionization chamber were performed using these phantoms to gauge their relative impact on outputs for 6, 9, 12, 16, and 20MeV electron beams from a Varian TrueBeam with cone sizes ranging from 6 × 6 cm2 to 25 × 25 cm2. The corresponding Monte Carlo simulations of the measured geometries were carried out using the CT scans of these phantoms. The results indicated that the Eclipse eMC algorithm can predict these output changes within 3% for most scenarios. However, at the lowest energy, the discrepancy was the greatest, up to 6%. In comparison, the Eclipse GGPB algorithm had much worse agreement, with discrepancies up to 17% at the lowest energies.

Figure 1

Phantoms with varied topology generated by adding a mass of water‐equivalent material at the surface of a Solid Water phantom. Added phantom masses had a height of 1.5 cm from the top of the Solid Water and were marked with a cross to indicate the isocenter.

Figure 2

3D rendering of the ten phantoms as seen in the beam's eye view and with different perspective.

Figure 3

Mean errors in percentage (averaged over all ten phantoms) between the correction factors ${\text{CF}}_{\text{topo}}$ from Eclipse eMC calculations and the measurements. The mean errors are also presented for the GGPB algorithm for comparison. The standard deviations on the means are depicted with the error bars. Energies ranging from 6 MeV to 20 MeV, SSD 100 cm, and $10\times 10\hspace{0.17em}{\text{cm}}^2$ cone used.

Figure 4

CF_topo factors as measured on the linac and predicted with the Eclipse eMC and the GGPB algorithms for 6 MeV electron beam and ten different phantoms.

Figure 5

CF_topo factors as measured on the linac and predicted with the Eclipse eMC and the GGPB algorithms for 9 MeV electron beam and ten different phantoms.

Figure 6

CF_topo factors as measured on the linac and predicted with the Eclipse eMC and the GGPB algorithms for 12 MeV electron beam and ten different phantoms.

Figure 7

CF_topo factors as measured on the linac and predicted with the Eclipse eMC and the GGPB algorithms for 16 MeV electron beam and ten different phantoms.

Figure 8

CF_topo factors as measured on the linac and predicted with the Eclipse eMC and the GGPB algorithms for 20 MeV electron beam and ten different phantoms.