Session Item

Currently, the IAEA TRS-398 Code of Practice is being updated. The updated version will provide values of beam quality correction factors (k_Q) for air-filled ionization chambers in clinical proton beams based on Monte Carlo simulations and experimental data. In recent years, the Monte Carlo codes PENH, GEANT4 and FLUKA have been investigated in terms of their feasibility for dosimetric calculations. It was shown that all three codes can be used to calculate k_Q factors to be in agreement with experimentally determined ones on the 1%-level.

In this study, we review the available data on Monte Carlo calculated f_Q factors in clinical proton beams which enter into Monte Carlo calculated k_Q factors. We provide average f_Q factors and discuss type-B uncertainties.

From f_Q factors as published in the literature, weight-averaged f_Q factors for different ionization chamber models were derived. The weight of each f_Q factor was indirectly proportional to its corresponding relative type-A uncertainty. In general, the amount of available data is not the same for all Monte Carlo codes. To avoid a code-specific bias, the weight of each f_Q factor from one Monte Carlo code was reduced if more than one value was available in the literature. Subsequently, the weight-averaged f_Q factors as a function of residual range (R_res) were fitted with a polynomial function of order 2.

In order to estimate the type-B uncertainty of weight-averaged f_Q factors, a rectangular distribution was assumed. The width of this distribution was defined as the difference between the largest and smallest f_Q factor enlarged by their corresponding relative type-A uncertainties.

In Figure 1 the fits of weight-averaged f_Q factors are shown in solid lines as a function of R_res for an exemplary cylindrical and plane-parallel ionization chamber. The uncertainty is indicated as dashed lines. Additionally, the original f_Q factors from the literature are depicted. For the cylindrical chamber f_Q decreases by 0.5% with increasing R_res. For  the plane-parallel chamber it is almost constant with variations of 0.2%.

Additionally, it can be seen that the agreement between the individual Monte Carlo codes is better for small values of R_res while the codes tend to diverge for larger values of R_res. Correspondingly, the uncertainty of weight-averaged f_Q factors increases with R_res and reach up to 2% (k=1), following the estimation described above.

Weight-averaged f_Q factors were derived from currently published Monte Carlo calculated f_Q factors. The Monte Carlo codes show a better agreement for small values of R_res while the codes diverge for larger values of R_res. This might be due to differences in the modelling of nuclear interactions whereas the role of nuclear interactions increases with energy and hence R_res. As a result, overall uncertainties of Monte Carlo calculated f_Q factors can be expected to be larger for higher energies.