ESTRO 37 Abstract book

S965

ESTRO 37

EP-1796 Template-based commissioning model for in vivo EPID transit dosimetry K. Landheer 1 , R. Van Oers 1 , I. Olaciregui-Ruiz 1 , A. Mans 1 1 Netherlands Cancer Institute, Radiotherapy, Amsterdam, The Netherlands Purpose or Objective For in-vivo dose verification, a back-projection (BP) model was applied on measured images using an electronic portal imaging device (EPID), which was positioned behind the patient. This model was calibrated for each combination of photon beam energy, linear accelerator (linac) and EPID. In this study, we compared the full commissioning method (FCM) and a novel template based commissioning method (TBCM). With the FCM, the complete parameter set of the BP model was created for each EPID. For the TBCM, a template parameter set (TCM) with non-EPID specific parameters was used and only EPID specific parameters were newly created. The commissioning workload per model was 2 h for FCM and 10 min for TBCM. Material and Methods We found our BP model parameters using an iterative fitting procedure, where EPID reconstructed dose distributions were fitted to ionization chamber (IC) values. A novel photon beam specific TCM was established with a fitting that uses an improved BP algorithm, including dose rate correction, and an input of a large pool of data from 4 different EPIDs. Previously, a TCM was created with the non-EPID specific parameters of only one FCM. We created 12 TBCMs and 12 FCMs for 6 EPIDs attached to 6 SL20i linacs with 6 and 10 MV beams. The model fits were assessed with the percentage dose difference between reconstructed and IC dose at the isocenter (ΔD isoc (%)) for a series of square fields from 3x3 to 23x23 cm 2 on a phantom. Furthermore, TBCMs and FCMs were used to reconstruct dose for 134 clinical treatment plans consisting of EPID data of 782 IMRT beams and 533 VMAT arcs. Planned and reconstructed dose were compared with the dose difference at the dose reconstruction point (ΔD DRP (%)) and the gamma (γ) evaluation parameters with (3 %, 3 mm): γ pass rate (%γ < 1), γ mean and the maximum 1% γ. Results The largest difference in the mean ΔD isoc (%)(TBCM – FCM) for the field size series was 1.06% (6 MV series B4a). ΔD isoc (%)(TBCM – FCM) was positive (compare Fig 1a with b, and c with d), since a dose rate correction was used in the BP algorithm for the fit of the TCM and not for the FCM. Also, the σ of the TBCM series was equal or a little more than for the FCM series.

The results for treatment plans in Table 1 show that the sigma of ΔD DRP (%) TBCM was equal (10 MV) or slightly larger (6 MV) than ΔD DRP (%) FCM. Also, mean ΔD DRP (%)(TBCM-FCM) was 1.07% (6MV) and 0.11% (10MV). The differences in the average γ pass rate, γ mean and γ 1% for the TBCMs and FCMs were small.

Conclusion The largest difference in the mean ΔD DRP (%) between TBCMs and FCMs observed with the treatment plans was 1.07%, and for the square field validation series, it was 1.06%. These dose differences and the differences in the gamma evaluation results were clinically acceptable. Moreover, the dose differences were mainly attributed to a correction in the BP algorithm used for fitting, which results in a reconstruction dose closer to the IC values. Thus, TBCMs were found to be clinically acceptable. EP-1797 Integrating an in vivo EPID dosimetry system into the Varian ARIA environment using scripting D. Welsh 1 , D. Burns 2 , F. Cutanda 1 , A. Sankar 3 1 Western General Hospital, Oncology Physics, Edinburgh, United Kingdom 2 Western General Hospital, Therapuetic Radiography, Edinburgh, United Kingdom 3 Ninewells Hospital, Radiotherapy Physics, Dundee, United Kingdom