Program Information
Assessment of Monte Carlo Simulations as Alternative Reference for Patient Specific Quality Assurance in Pencil Beam Scanning Proton Therapy
J Shin*, K Jee , B Clasie , N Depauw , T Madden , H Paganetti , H Kooy , Massachusetts General Hospital, Boston, MA
Presentations
SU-H1-GePD-T-5 (Sunday, July 30, 2017) 3:00 PM - 3:30 PM Room: Therapy ePoster Lounge
Purpose: In our clinical workflow, two 2D dose measurements per pencil beam scanning (PBS) field are performed at shallow depth (SD) and tumor depth (TD) with patient-specific devices. Our purpose was to assess whether patient-field-specific quality assurance (QA) measurements can be replaced by using an accurately commissioned Monte Carlo (MC) tool. The feasibility was investigated through γ-index analyses.
Methods: A total of 100 field measurements (depths range from 2 cm to 25 cm) from 13 prostate plans were chosen and simulated by our TOPAS/MC-based QA framework, which was validated within ±2 % in absolute dose level over a whole treatment range and within ± 1 % against a reference QA field. The simulations were designed to achieve 2 % statistical uncertainty inside the penumbra within 5 to 6 hours per treatment plan. The 2D γ-index (3%/3mm with 10 % threshold) of analytical dose algorithm (planning system) versus measurement (A2D) and MC versus measurement (M2D) were analyzed as a function of depth relative to the isocenter. Furthermore, the γ-index of 3D dose in the water phantom from MC and analytical algorithm (A2M) was analyzed.
Results: Both A2D and M2D showed excellent passing rates (γ≤ 1.0) at TDs, e.g., the average passing rates were 98.3±2.8 % and 98.3±2.4 % respectively. However, the passing rate of A2D decreases with shallower depths while those of M2D remains consistent at shallow depths, 84.7±20.6 % and 97.6±3.5 % respectively for SDs less than 0.75. The average and median passing rates of A2M were 97.9±1.4 % and 98.3 %.
Conclusion: This study experimentally confirmed that for patient-specific QA, MC shows high passing rates independent of the depths while analytical algorithms exhibit pronounced dose discrepancies at shallow depths possibly due to inaccurate modeling of slit-scattering. The use of MC as a reference in patient-specific QA is thus recommended.
Contact Email: