The prediction of the β+-activity distribution on the basis of the patient CT using the FLUKA simulation code

The in-beam PET technique for quality assurance of C therapy is based on the comparison of β-activity distributions measured during therapeutic irradiations with those predicted from the treatment plan. The generalisation of this method to all species of therapeutically relevant ions requires a precise description of both the stopping process and the nuclear interactions of the projectiles in tissue. The fluka radiation transport simulation code [1], [2] may meet these requirements since the implementation of suitable models for describing light ion nuclear interactions is in progress. Furthermore, it has been demonstrated that measurements of proton beam induced β-activity distributions can be rather well reproduced by calculations based on the fluka code [3]. The approach developed for the proton case may be feasible for other ion species: In a first step the spatial distribution of the flux of primary and secondary particles φ(E,A,Z, r) is calculated on the basis of the fluka internal models of nuclear reactions. In a second step a realistic description of the production of positron emitters is obtained by combining these results with measured or semi-empirical cross sections [3]. The β-activity produced via target fragmentation strongly depends on the stoichiometric tissue composition, especially on the O/C ratio, which varies between 0.22 for adipose tissue and 4.05 for muscle. This ratio determines the relative abundances of O (t1/2 = 2.03 min) and C (t1/2 = 20.38 min) as the dominating β-active reaction products, and because of the different half-lives the activity in different tissue types. Especially for projectiles with Z < 5, where β-emitter can only be produced by target fragmentation taking into account the tissue stoichiometry is unavoidable for a correct description but also for heavier projectiles as e.g. C a refinement of the prediction of the β-activity distribution from the treatment plan is expected. In a typical patient CT the Hounsfield numbers range from −1024 to 3071 representing different densities and different chemical compositions of tissue. The CT numbers are converted by a Fortran code into a file which is