The final aperture superposition technique (FAST) is a method to reproduce rapidly the electron-beam depth dose curves and output factors that would be calculated by a full Monte Carlo simulation. FAST uses precalculated Monte Carlo-based differential dose arrays and performs a superposition of open and shielded contributions to account for arbitrarily shaped insert openings. The objective of this work was to refine and validate the accuracy of the FAST method for a full range of treatment parameters. Compared to full simulations, raw FAST calculations tended to underestimate dose near the surface deposited by particles that crossed the shield-opening interface of the insert. In this study, a set of empirical correction curves was derived to reduce the errors from this "collimator effect." FAST and full simulation calculations were compared for every combination of six beam energies (6-21 MeV), four applicator sizes (10-25 cm), and two source-to-surface distances (SSDs) (100 and 110 cm). Validation tests were performed for a total of 192 fields using four sample insert openings: an open insert and 2, 3, and 5 cm diameter circular openings. Calculations were also performed for four patient inserts with irregularly shaped openings. Using the empirical correction curves, systematic errors were reduced, resulting in mean dose differences of less than 1% of the maximum full simulation dose. FAST relative output factors reproduced full simulation output factors to within 3% for all configurations except for the 2 and 3 cm diameter openings for the 6 and 9 MeV beams at 110 cm SSD. The maximum shift between the FAST and full simulation depth dose curves in the 90%-80% fall-off region was less than 3 mm for 97% of the fields. For the patient insert calculations, differences in output factors and mean differences in depth dose curves were within 1.5% with maximum shifts of 1.5 mm in the 90%-80% fall-off region. A small set measurements also demonstrated 3% accuracy in FAST output factors except for a 5% deviation for a 2 cm diameter insert for the 6 MeV beam at 110 cm SSD. These results demonstrate that FAST can be used to provide output factors and depth dose curves for most clinical cases.

No abstract

Purpose: To quickly calculate central‐axis depth doses for arbitrarily‐shaped electron inserts using precalculated differential dose data. Method and Materials: Monte Carlo simulation (EGS4) is used to obtain phase space distributions in a plane upstream from the insert. From these distributions, central‐axis dose in a water phantom is calculated for a) an open configuration with no insert in the applicator and b) a fully shielded configuration with a completely solid insert. Differential dose datasets are created by tagging the location of particles as they pass through the insert and scoring contributions to the dose from different regions. For arbitrarily‐shaped inserts, the Final Aperture Superposition Technique (FAST) code adds open and shielded contributions given an outline of the insert opening. Differential dose datasets are calculated for a range of energies (6–21 MeV), applicator sizes (10–25 cm), and source‐to‐surface distances (100–110 cm). FAST calculations are compared with full Monte Carlo simulation through the insert using the same initial phase space distributions. Results: FAST depth doses are similar to the full simulation but tend to be slightly lower in dose, particularly for higher energy beams. This “collimator effect” is due to the inability of FAST to score dose from particles that escape the shield through the edges of the insert opening. An average collimator effect curve has been defined using a linear fit to the average difference between full simulation and FAST for a 2 cm diameter opening and a fully open applicator. Using this curve to correct the FAST doses results in mean residual errors and standard deviations in these errors of less than a percent. Conclusion: With a minor correction for the collimator effect, the FAST program can accurately calculate central‐axis doses for a large range of treatment parameters.