Unencrypted login | home

Program Information

Verification of Acuros Dose Calculation Accuracy in Lung SBRT


L Wang

L Wang*, K Bush, L Xing, E Mok, Stanford Univ School of Medicine, Stanford, CA.

SU-E-T-531 Sunday 3:00:00 PM - 6:00:00 PM Room: Exhibit Hall

Purpose: The AcurosXB Advanced Dose algorithm (Varian Medical Systems) was reported to give more accurate dose calculation in heterogeneous medium compared with the AAA algorithm. In this work, a validation of AcurosXB dose calculation accuracy is performed for use in lung SBRT.

Methods: A CIRS (Computerized Imaging Reference Systems, Inc.) lung phantom was used with a tissue equivalent 2cm circular target inserted into the lung. The phantom was customized to hold a Gafchromic EBT2 film through the center of the target. The phantom was CT scanned, and single field (3cm by 3cm anterior/lateral) plans, as well as VMAT plans were created for 6 and 10 MV energies with different delivery modalities (with/without a flattening filter). All plans delivered 500 cGy to the PTV which was the target plus 0.5 cm uniform margin. Plans were delivered on a Varian TrueBeamTM STx using image guidance to locate the isocenter. The film dose was compared with Eclipse v10.0 dose calculated with AAA and Acuros algorithms using FilmQA software (3cognition). Monte Carlo simulations (BEAMnrc/DOSXYZnrc) were performed to further validate AcurosXB calculations.

Results: Within the GTV, AAA and Acuros algorithms were in good agreement with the film measurements to within 2% mean dose difference. AAA calculation consistently overestimated the dose to the lung surrounding the GTV by 5-9% for the 10MV beam and 2-4% for the 6MV beam. Acuros and MC calculations were found to be in good agreement with EBT2 film measurement within 2% at 6MV including the build up and build down regions.

Conclusions: The CIRS customized lung phantom with EBT2 films was found to be an excellent tool in validating the dose calculation algorithm for lung SBRT application. AcurosXB was found to be more accurate than AAA algorithm in lung and near lung/tissue interfaces.

Contact Email