EPID-Based in Vivo Dosimetry: The Next Step in Patient-Specific QA
B Mijnheer*, I Olaciregui-Ruiz, R Rozendaal, J-J Sonke, H Spreeuw, R Tielenburg, M van Herk, A Mans, The Netherlands Cancer Institute, Amsterdam, NetherlandsTU-E-108-8 Tuesday 2:00PM - 3:50PM Room: 108
Purpose: To demonstrate that EPID-based in vivo dosimetry can play an important role in the total chain of verification procedures in a radiotherapy department.
Methods:In our institution a back-projection algorithm has been implemented for the in vivo dose verification of all treatments, including IMRT and VMAT, using a-Si EPIDs. For non-IMRT plans the reconstructed dose is compared for each beam with the planned dose in a plane parallel to the EPID intersecting the isocenter. For IMRT and VMAT the total 3D dose distribution is reconstructed and compared with the planned dose distribution using 3D gamma evaluation. Dose-reconstruction and gamma-evaluation software runs automatically in the clinic; i.e. no user intervention is required for the analysis. Deviations outside the alert criteria are investigated by a medical physicist with respect to their clinical relevance. If necessary, additional phantom measurements are performed to separate patient-related errors from planning or delivery errors.
Results:With the current set of alert criteria, 63% of 3,839 patient treatments in 2012 were automatically approved without any human intervention. Most deviations were due to limitations of our back-projection and dose evaluation software and could easily be recognized. Other sources of error were patient related: anatomical changes or deviations from the routine clinical procedure, and would not have been detected by pre-treatment verification. Inspection of cone-beam CT scans yielded information about anatomy changes such as shrinkage of the tumor and weight loss. Furthermore, 3D gamma-evaluation of in vivo dose distributions obtained during verification of VMAT head-and-neck treatments showed a high correlation with DVH parameters.
Conclusion:EPID-based in vivo dosimetry as applied in our institution provides clinically more useful information and is less time consuming than patient-specific pre-treatment dose verification. In our opinion, large scale implementation of EPID-based in vivo dosimetry is therefore the next step in patient-specific QA.