Experimental benchmarking of RayStation proton dose calculation algorithms inside and outside the target region in
Sirinya Ruangchan1, Barbara Knäusl2, Hermann Fuchs2
1Division of Medical Radiation Physics, Department of Radiation Oncology, Medical University of Vienna, Vienna, Austria; Division of Therapeutic Radiation and Oncology, Department of Radiology, King Chulalongkorn Memorial Hospital, Bangkok, Thailand.
Purpose:
The aim of the presented study was to complement existing literature on benchmarking proton dose by comparing dose calculations with experimental measurements in heterogeneous phantom. Points of interest inside and outside the target were considered to quantify the magnitude of calculation uncertainties in current and previous proton therapy practice that might especially have an impact on the dose in organs at risk (OARs).
Methods:
The RayStation treatment planning system (RaySearch Laboratories), offering two dose calculation algorithms for pencil beam scanning in proton therapy, i.e., Pencil Beam (PB) and Monte Carlo (MC), was utilized. Treatment plans for a target located behind the interface of the heterogeneous tissues were generated. Dose measurements within and behind the target were performed in a water phantom with embedded slabs of various tissue equivalent materials and 24 PinPoint ionization chambers (PTW). In total 12 test configurations encompassing two different target depths, oblique beam incidence of 30 degrees and range shifter, were considered.
Results:
PB and MC calculated doses agreed equally well with the measurements for all test geometries within the target, including the range shifter (mean dose differences ± 3%). Outside the target, the maximum dose difference of 9% (19%) was observed for MC (PB) for the oblique beam incidence and inserted range shifter.
Conclusion:
The accuracy of MC dose algorithm was superior compared to the PB algorithm, especially outside the target volumes. MC based dose calculation should therefore be preferred in treatment scenarios with heterogeneities, especially to reduce clinically relevant uncertainties for OARs.
