Validation of the Mobius system for patient-specific quality assurance using introduced intentional errors

Mobius3D and MobiusFX are model-based verification tools for treatment plan dose calculation and treatment delivery. The software facilitates patient-specific quality assurance by extracting data from linear accelerator treatment log files and performing a 3D dose calculation on the original patient CT dataset using an independent collapsed cone algorithm. In this study, we evaluate the ability of the Mobius system to detect linear accelerator-related errors compared with existing measurement-based systems, namely the ArcCHECK® and 3DVH® systems. Three original treatment plans and 47 plans with introduced delivery errors, for a total of 50 plan deliveries, were investigated. The results from this study demonstrated comparable gamma passing rates and error detectability between the Mobius and ArcCHECK® systems while the 3DVH system generally exhibited a lower sensitivity. This work also demonstrated the ability of the Mobius system to detect delivery errors of down to 2° collimator rotation, 1 mm MLC bank offset and 10 mm collimator jaw offset.