Computer simulations are a powerful tool for modelling and predicting complex phenomena in the natural world, where they model the environment, financial markets, and medicine. The application of simulations in the world of medical physics are evident, with examples such as modelling particle transport. In this research, a functioning simulation of a single-photon emission computed tomography system was developed as a tool for medical physicists to use for validation and experimentation. Computer simulations are particularly useful in medical physics as they allow for non-invasive experimentation without the potential for harm to any patients.

A collimator extension was successfully developed in the TOPAS C++ extension framework. The extension allows users to create hexagonal-holed collimators of any size, and allows modification of key parameters such as: material, hole length, hole diameter, and septal thickness in a way akin to definitions of other geometry components. The collimator behaviour was validated, showing that it successfully absorbed diagnostic energy photons at non-perpendicular angles, allowing for the resolution of sources of different shapes. Measuring collimator efficiency showed that efficiency was typically higher than expected, however this result was inconclusive. A digitiser element was developed to simulate the non-collimator elements of the gamma camera, namely the scintillation crystal, photomultiplier tubes, and associated electronics. The Python-based solution was able to match the behaviour of a real system in terms of energy resolution and windowing across multiple tests. The spatial blurring and resolution tests showed that the digitiser consistently over-blurred detected photons leading to unrealistically poor spatial resolution performance over a range of target resolutions in the typical system range. Finally, the capabilities of both sections integrated together were tested by simulating a myocardial perfusion imaging study. Sources simulating a normal and abnormal heart were imaged at a range of projection angles before then having a central slice processed into a sinogram and reconstructed. The projection images were transformed into sinograms and those were successfully reconstruction with both a backprojection and filtered backprojection method. Unfortunately, low photon counts in each projection image, especially in the abnormal heart simulation, led to poor quality reconstructions where it was not clear if a slice corresponded to the normal or abnormal heart source.