Charged particle radiobiology beamline using tandem accelerator-based MeV protons and carbon ions: a pilot study on the track-end radiation quality, variable biological effectiveness and Bayesian beam dosimetry.

For in vitro cell irradiation using tandem accelerator-based MeV protons and carbon ions, by TOPAS simulation, a pilot study of performance evaluation is presented on a collimation beamline for 3 MeV protons and 10 MeV carbon ions from a 2  ×  3 MV tandem accelerator. Based on the elements and source parameters, a collimated beam of 2.8 MeV protons or 2.5 MeV carbon ions, with 5.175 mm or 5.166 mm full width tenth maximum (FWTM), respectively, can be delivered to the target cell dish. TOPAS simulations and/or deterministic algorithms present a Bragg curve of linear energy transfer (LET) (10-70 keV μm) along a 138 μm range of the proton beam, and a declining LET of the carbon beam (900-100 keV μm) within 4 μm range. Based on the biophysical models for relative biological effectiveness (RBE) of protons, TOPAS RBE scorers presents a set of depth-variation curves of the proton RBE (for V79 and DU145 cells), linearly related to the Bragg curve of the proton LET. Based on the microdosimetric-kinetic (MK) theory, in the 4 μm range for a monolayer cell thickness, the mean RBEα (V79 cells) of the carbon ion beam is estimated as 3.612 (late S phase) and 1.737 (G /S phase) for the mean LET of 492 keV μm. For practical irradiations, a tunable proton RBE can be acquired by changing the thickness of the cell dish. For the low-energy high-fluence (rate) beams, indirect beam measurements are proposed to detect the proton-beam induced scattering/recoil protons from a beam-intercepting Mylar film, and the carbon-beam induced backscattered electrons from a gold-deposited Havar-foil beam window. Statistical dosimetry for the indirect measurement is established, using a Bayesian model based on the preset number of detection counts, by which the mean value of the whole-dish dose can be prescribed and the uncertainty introduced in the survival data can be corrected.