A System for Verifying Radiation Therapy Treatments That Use Multiple Modes of Real-Time Motion Adaptation.

This technical report describes the development and use of a system for performing routine quality assurance (QA) of real-time adaptive radiotherapy treatments delivered using the Accuray Radixact Synchrony platform. Established international guidelines recommend the use of a dynamic phantom to verify real-time target tracking and the use of 2D dosimetry to verify the accuracy of modulated radiotherapy treatment delivery; hence, a dynamic QA system was developed to fulfill both of these requirements while being comparatively easy to understand, consistent with other local radiotherapy QA processes and straightforward to set up and use. The resulting dynamic QA system used a PTW Octavius 1500 ionisation chamber array dosimeter, set up in the bottom half of an Octavius II phantom, on top of a CIRS Dynamic Platform that enables superior-inferior motion, with purpose-designed 3D-printed components placed on top of the array. The 3D-printed components included lung-equivalent blocks containing a unit-density sphere, for testing Synchrony's lung tracking with respiratory modelling mode, and five implanted gold seeds, for testing Synchrony's fiducial tracking mode. The use of the dynamic QA system for verifying the performance of Radixact Synchrony's fiducial tracking mode was demonstrated by measuring the dose from eight prostate treatments, while four different sample prostate motion traces were used to drive the dynamic platform. The results of this testing indicated that Synchrony was able to track the motion of the platform accurately within 1 mm (often within 0.5 mm), and local patient-specific QA gamma evaluation thresholds were consistently surpassed for all motion traces other than a high-frequency prostate motion trace. For this challenging high-frequency motion, gamma agreement indices from QA measurement results decreased by more than 3.5% when Synchrony was not used; without Synchrony's real-time tracking, the target so often moved away from its planned location that substantial proportions of the treatments were delivered at unintended locations. The use of the dynamic QA system to verify the performance of Radixact Synchrony's lung tracking and respiratory modelling mode was demonstrated by tracking the 3D printed unit-density sphere while the phantom moved according to a pre-programmed sinusoidal waveform. Again, the resulting 2D measurement exceeded the local patient-specific QA gamma evaluation threshold. This result, in combination with the fiducial tracking results, suggests that the dynamic QA system developed in this work would also be useful for verifying Synchrony's performance in fiducial tracking with respiratory modelling mode, for abdominal targets. Combined testing of Radixact jaw and multileaf-collimator motions modified by Synchrony may be achieved by applying a yaw rotation to the dynamic platform so that lateral and superior-inferior phantom motions are produced. The dynamic QA system described in this technical report is well suited to routine clinical verification of tracking modes provided by Radixact Synchrony (and other real-time adaptive radiotherapy systems) during initial commissioning, periodic end-to-end testing, and routine patient-specific QA testing.