Bremsstrahlung imaging has been used for targeted radionuclide therapy to assess the distribution of beta-emitting isotopes within a region. This enables accurate dose estimates to be made. Monte Carlo techniques have been used to simulate the radiation processes involved to allow the technique to be optimized. However, if ⁹⁰Y is used as the nuclide, only 8% of the beta decays eventually result in the emission of a bremsstrahlung, so the simulation process is inefficient and computationally expensive.
To speed up the simulations, it is proposed that for a point source the beta emission step is disregarded and the point beta source is replaced with a distributed gamma source that replicates the distribution of photons produced with respect to the distance from the source, the spectrum of the emissions and the angular distribution of the emissions. The distributed source is termed a ‘kernel’.
10⁹disintegrations of a point ⁹⁰source in water were initially simulated using GEANT4. This allowed a data base of the resulting bremsstrahlung emissions to be obtained. The kernel was then constructed as a series of concentric shells from which the bremsstrahlung are emitted. For each shell, photons are emitted with the spectrum and angular distribution consistent with that range. It was found that the kernel was accurate if the shells up to 5 mm from the source had incremental radii of 0.1 mm, then 0.5 mm increments out to 7 mm and 1 mm increments beyond that. The kernel reduced simulation times by a factor of 27.