Evaluation of EGS4/PRESTA multiple-scattering algorithms for 90Sr/90Y intravascular brachytherapy dosimetry

The purpose of this work is to evaluate the EGS4/PRESTA electron multiple-scattering (MS) algorithms for dose calculation in intravascular brachytherapy (IVBT) using a 90Sr/90Y source. The small source size and the small volume of interest in IVBT require very fine spatial resolution, which may break down the constraints of Moliere's MS theory as implemented in EGS4. The theory is accurate only when the electron step sizes are large enough to allow the number of collisions Ω0 to be much greater than e = 2.7183. When step sizes are too small to allow at least 2.7183 collisions, as may be necessitated by the fine geometry, the algorithm may switch off MS, producing dosimetric artefacts. This study showed that switching off MS could produce a dose deviation of up to 6% when the half-thickness (d/2) of the dose scoring region is comparable with the Moliere minimum step size (tmin = 2.7183). The effect of switching off MS is negligible if d/2>>tmin. For the case of Ω0>e, if the electron step sizes are chosen to allow five to 40 collisions, with increasing step size, the doses surrounding the source increase and the error decreases. On the other hand, when larger step sizes are chosen, the dose calculation voxel size must also be increased in order for the calculations to converge. A good compromise between accuracy and applicability for IVBT simulation can be made, if the thickness of the scoring region is 0.1 mm and the electron step sizes are in the range allowing 10 to 30 collisions. (author)