SU-E-T-538: Validation of Monte Carlo Dose Algorithm in Heterogeneous Medium

Abstract

Purpose: Evaluate validity of Monte Carlo(MC) based treatment planning system(TPS) used for lung SBRT. Methods: A commercial TPS(XVMC, iPlan4.1, BrainLAB, Germany) was validated using benchmark dose measurements in water and heterogeneous‐slab phantoms. Five phantoms were built using four different density materials: tissue‐equivalent plastic‐water, lung‐equivalent(low‐& high‐density) cork, and bone. EDR‐films at various depths in phantoms were irradiated with 6MV photons. MC plans were calculated dose to medium using 2mm grid‐size and 1% dose‐variance. Ion chamber and film measurements of depth‐dose, dose‐profiles and output‐factors for a range of field‐sizes(12×12 – 60×60mm) were performed. Measured versus calculated dose profile differences were quantified using two dose‐indices: Ddiff for the central 80% of field, and Dspill for dose outside field‐edge(50%−10% dose). Results: For homogeneous and hetero‐bone phantoms, MC and PB calculations agreed well with measurements. For hetero‐lung phantoms, there was excellent agreement between MC‐calculations and measured doses (Ddiff <3%) for all field‐sizes and depths. The agreement improved with increasing field‐size(2.6% for 12×12mm vs. 1.1% for 60×60mm @4cm depth in lung). However, at the same depth and field‐size, PB calculations significantly over‐predicted measured dose (34% and 6.7% respectively). PB and measured dose differences increased with decreasing field‐size, decreasing density and increasing depth within heterogeneity, but converged to <1% beyond heterogeneity. In contrast, significant under‐prediction of dose by PB model(up to 50%) was seen in the penumbra region outside field. Dspill differences between calculated and measured dose increased with field size but were lower for MC than PB. Conclusion: Our TPS validation measurements highlight need for MC planning in lung SBRT. PB calculations Result in significant under‐dosing of target and over‐dosing of surrounding tissue. Over‐prediction of target dose is inversely proportional to field‐size. Knowing the magnitude of Ddiff for PB for a given field‐size, one can easily calculate this index at any other field‐size.