Stereotactic Ablative Body Radiotherapy (SABR) is a type of treatment modality that is used to treat cancer. SABR uses high dose radiation beams and is delivered in a lower number of treatments compared to conventional radiotherapy. Since SABR uses high doses per treatment with a low number of treatments, errors in the beam parameters can potentially have significant dosimetric effects on dose delivered to the patient. In this research, the efficacy of Python-based models to simulate linear accelerator errors were studied as Python models can be performed much faster than Monte Carlo simulations, which is advantageous for simulating errors quickly in the clinic. To compare the validity of these results, the same errors were modelled in a Monte Carlo-based software called TOPAS MC which was considered as the gold standard for the purposes of this research. The Python and TOPAS models were compared by using dose difference between dose distributions, correlation between Python and TOPAS dose distributions, and the gamma percentage pass rates. It was found that Python could sufficiently reproduce the TOPAS linac errors well for collimator and SAD error but not for MLC translations and couch pitch errors. Reproducing combinations of introduced errors in Python resulted in varying degrees of success, depending on the specific error combinations.

This research shows that Python can be used as a viable alternative to model specific linear accelerator errors provided the error is small (+1 mm, +1.5° deviations) and all other components are working perfectly. To supplement this research, further research needs to be done to better model certain errors in the linear accelerator as well as understanding the range for which these errors are valid.