Transit Dosimetry for VMAT Online Adaptive Radiotherapy with Monte Carlo-Generated Prediction Images
Abstract
Purpose
To develop a Monte Carlo (MC)-based approach for generating transit dosimetry prediction images for volumetric modulated arc therapy (VMAT) online adaptive radiotherapy (OART) on a ring gantry linac featuring an advanced cone beam computed tomography (CBCT) imaging system.
Methods
A frame grabber saves frames acquired by the electronic portal imaging device (EPID), which measures the radiation exiting through the patient during OART treatment delivery. MC simulations are used to generate transit dosimetry prediction images, based on a previously validated computational model of the linac, patient, and EPID. The EGSnrc MC software is used. The phantom/patient model is based on the CBCT acquired at the beginning of the OART session. This CBCT is also used for OART treatment planning. Transit dose frames can be analyzed individually or in groupings of arbitrary size. An anthropomorphic pelvis phantom and a patient test case are considered (both prostate VMAT, 8 Gy per fraction).
Results
For the VMAT plan delivered to the pelvis phantom, CBCT reconstruction technique (with and without iterative reconstruction and scatter correction) had a negligible effect on MC-based transit dose prediction images when considering 30-degree gantry angle segments (~175 frames per segment). The human patient test case shows gamma pass rates ≥ 95% for each VMAT field.
Conclusion
This study demonstrates the feasibility of transit dosimetry during OART, using a frame grabber to allow for a gantry-angle-resolved analysis. MC simulations based on the OART CBCT and plan-of-the-day create transit dosimetry prediction images that can be compared with EPID measurements. This approach can be used to ensure that the treatment was delivered as intended, and has the potential to identify treatment delivery errors due to e.g., patient-related changes or machine malfunction.