Purpose: To assess the clinical impact of Monte Carlo (MC) dose calculations in proton therapy by analyzing uncertainties of analytical algorithms when predicting dose to target and critical structures and beam ranges in patient geometries. Methods: We used TOPAS to simulate double scattering proton treatments, which were compared to analytical treatment planning dose calculations (TPC). We investigated: 1)beam ranges, dose distributions, does‐rate profiles and output factors using various dedicated experiments.range differences caused by the patient heterogeneity considering 48 fields and 4 treatment sites. We calculated distal 2)dose surfaces composed of beam ranges and assessed average range difference (dR) and root‐mean‐square deviation (RMSD).3)dose delivery accuracy using a DVH based analysis considering patient treatment plans from 3 sites. 4)the accuracy of dose delivery for small fields (diameter below 7cm) based on patient heterogeneity for 38 head fields. We assess differences in D50. Results: 1)MC and measurements agree within statistical uncertainties.2)For 3 of 4 patient sites the RMSD between TPC and TOPAS was <2mm. Head treatments indicate significant RMSD of up to 6mm. 3)TPC DVHs showed underdosage for head cases by 2% compared to TOPAS, discrepancies in critical volumes can reach up to 50%, prostate patients show significant penumbra differences, liver patients show good agreement between TOPAS and TPC. The site‐specific findings can be explained by beam range and geometrical complexities.4)For small proton fields, discrepancies of up to 5.4% were found in D50 for single fields. A clear correlation between the accuracy and patient heterogeneity was established. Conclusion: This study highlights treatment sites and scenarios where MC simulations are recommended for proton therapy, i.e. small fields (due to scattering disequilibrium) and head treatments with bone/air/tissue interfaces affecting the ability of analytical algorithms to predict the correct beam range. Absolute dose differences where typically within 2% for large fields.
All Science Journal Classification (ASJC) codes
- Radiology Nuclear Medicine and imaging