ESTRO 37 Abstract book

S1138

ESTRO 37

Material and Methods Forty three prostate cancer patients, treated with 77 Gy in 35 fractions, were included in this retrospective study. At each fraction the position of patient is corrected based on the registration of pre-treatment cone beam CT onto the pCT according to the bony anatomy. Although the target group do not have markers, we have used patients who have fiducial markers implemented in the prostate for the purpose of this analysis. Offline fiducial marker registrations were performed on all cone beam CT scans to establish the prostate translational errors. The cone beam images of first few initial fractions have been used to calculate the average prostate motion based on residual error between bony anatomy match and marker match. The average prostate motion is subtracted from setup error of remnant fractions if the displacement exceeded a threshold value. We changed the threshold from zero to five millimeters. The systematic residual prostate motion of whole treatment fractions is calculated. The number of initial cone beam scans used to calculate the average prostate motion is varied to find the optimal moment to adapt the treatment plan. The workload was estimated based on the chosen value of threshold. In clinical practice the initial treatment plan can be adapted at fraction N+1 based on setup error of the first N fractions. Results The minimum residual systematic error is achieved by correcting based on the first 13 fractions, figure 1. The reduction in residual systematic prostate motion was more profound in CC and AP directions. A correction with zero action level, based on the data of the cone beam scan of first 13 fractions reduced the systematic error by a factor of 2. The group systematic errors increased when the prostate motion during the first fraction was used for correcting. This suggests that initial fractions might not be representative of prostate motion during the whole treatment course. Figure 2 shows the percentage of patients who need a replanning versus the chosen threshold value. Selecting a threshold of 2mm led to 88% of patients requiring a replanning.

applied to an FF beam the parametrization will naturally yield values close to those already reported in literature. Conclusion The novel parametrization works on general FF and FFF beams. Using it we were able to characterize the MR- Linac beam penumbra without reference beam. Upon upgrade to the clinical version of the MR-Linac, we will confirm the penumbra values with a new set of measurements.

EP-2072 Reducing Systematic Error For Radiotherapy Of Prostate And Pelvic Lymph Nodes S. Nakhaee 1 , M. Buijs 1 , P. Remeijer 1 1 Netherlands Cancer Institute, Radiotherpay, Amsterdam, The Netherlands Purpose or Objective The prostate can move relative to bony anatomy due to changes in bladder or rectal filling. Our current approach to account for daily setup errors and anatomical variations is image-guided re-aligning of patient bony anatomy. However, the changes in relative positions of prostate and bony anatomy cannot be corrected by only repositioning the patient. In this study we ran a simulation to quantify the reduction of systematic residual prostate motion by an adaptive procedure which utilizes the average prostate motion during the initial fractions.