Room: AAPM ePoster Library
Purpose: recent MR-LINAC system allows for daily MR image guided on-line adaptive radiotherapy. The on-line adapted plan must be verified before beam delivery, while current measurement based QA is not feasible. It is the purpose of this work to investigate the independent dose verification based on a fast commercial convolution/superposition (C/S) dose engine for MR guided on-line plan QA.
Methods: third-party treatment planning system (refer as TPS_QA) with C/S dose engine for photon beam dose calculation without consideration of magnetic field was employed in this work. Dose distributions from different field size were calculated with clinic MR-LINAC TPS (refer as TPS_clinic) for beam modeling. Attenuation with couch and cryostat were also finely tuned to match the calculation from TPS_clinic. After beam modeling, several test plans from TG-119 and 30 clinic plans were calculated on both systems and compared with either point dose measurements on solid water phantom or ArcCheck measurements.
Results: magnetic field causes electron return effect, which moves the maximum dose point backward and shifts the penumbra, while C/S algorithm without magnetic field can not model this effect and therefore shows a large dose calculation error in build-up and penumbra regions. For point dose comparison on TG-119 plans, TPS_clinic shows a dose deviation less than 2% with measurement, while TPS_QA shows a larger dose deviation up to 5% with measurement. For 30 patients dosimetric criteria comparison, the difference between mean PTV dose shows a most stable result of less than 2% of prescription dose for all patients.
Conclusion: C/S dose engine can be used for quick on-line plan dose check, the difference of mean PTV dose is stable and can be used as the simple verification criteria.
Funding Support, Disclosures, and Conflict of Interest: Natural Science Foundation of Guangdong, China (No. 2018A0303100020) and National Natural Science Foundation of China (No. 11805292)
Not Applicable / None Entered.