Monte Carlo Simulation of /sup 90/Y Bremsstrahlung Imaging

Abstract

Bremsstrahlung imaging can be used for the dosimetry of radionuclide therapies with pure beta emitters such as /sup 90/Y. However, the acquired spectrum is complex and continuous, making it difficult to select the optimum collimator and energy window settings and to perform scatter correction. This work used Monte Carlo simulation of the acquired spectra and point spread functions (PSFs) to investigate why certain acquisition parameters, usually established by experiment, are more effective than others. The aim was to establish a theoretical framework for their selection. An ADAC Forte gamma camera was simulated, comprising of collimator, NaI scintillator and backscatter compartment. Gaussian blurring was used to simulate energy resolution and intrinsic spatial resolution. Photon kernels were derived as substitutes for beta point sources to overcome the slow simulation of charged particle transport and bremsstrahlung production. Simulation speeds were improved by a factor of 200. Point sources of /sup 90/Y at various depths in water and with various collimators were simulated, and the type and location of photon interactions before detection were recorded. Resultant spectra and PSFs were analysed to establish optimal imaging parameters. Primary photons were the smallest spectral component at all energies. The spectrum was dominated by object scatter and camera interactions (lead X-rays, penetration and backscatter) in proportions which varied with energy window and collimator type. The highest image contrast was achieved with medium energy collimation at 100-150 keV. Scatter response functions derived from the PSFs offer potential for image correction.