Microdosimetry of a clinical carbon-ion pencil beam at MedAustron, Part 2: Monte Carlo simulation.

Abstract

Context: A first benchmark of the Monte Carlo (MC) code GATE was performed for microdosimetry in carbon ion beams. A validated MC model can, in turn, be used to calculate the radiation quality in lineal energy for any number of energies or treatment plans.

Materials and methods: Measurements were performed at five depths along a carbon ion pencil beam with a nominal energy of 284.7 MeV/u, with additional offside central axis (off-CAX) measurements at four depths. The silicon-on-insulator detector used was modeled with a simplified sensitive volume geometry in the Monte Carlo (MC) toolkit GATE. The source code of GATE was modified to allow the scoring of lineal energy for slab sensitive volumes.

Results: On average, the difference between the measured and simulated spectra (assuming the same keVµm^-1 cut-off value) was 15 % and 13 % for the frequency-mean and dose-mean lineal energy, respectively. By applying a shift in depth of 500µm towards the beam nozzle, the differences decreased to 10 % and 5 %, mostly affecting the positions in and the fall-off after the Bragg peak. The position in the fragmentation tail showed an edge at a different position than the expected carbon ion edge, which was determined through theoretical (stopping power tables) as well as computational (MC) means to be caused by boron ions.

Conclusion: MC is a powerful tool for any potential future clinical application of microdosimetry, provided the beam model has been benchmarked with experimental data. The detector geometry can be approximated with its sensitive volume if the water equivalent thickness of the detector is well known.