ESTRO 38 Abstract book

S154 ESTRO 38

converges towards the final motion-including DVH. The accuracy of DoseTracker’s real-time calculated DVHs was first investigated by simulating 39 fractions from 13 liver SBRT patients with Calypso-measured motion. The motion-induced reduction in the CTV D95% (ΔD95%) after each fraction was compared between DoseTracker and our treatment planning system (TPS; Eclipse). Next, the real- time DVH concept was used in two simulated dose-guided treatment scenarios where inter-field couch corrections were performed to correct mean position errors if ΔD95% exceeded either 5% or 10%. Results Fig 1A compares the motion-including DVH(t) with the planned static DVH at different time points during a treatment. DoseTracker’s final real-time estimated ΔD95% was in general in good agreement with the TPS with a root- mean-square error of 2.3 %-points and with the largest errors for small tumors (Fig 1B).

Conclusion A new method for real-time motion-including dose evaluation that relates to a traditional DVH was presented and used for simulated dose-guided couch corrections based on reconstructed tumor dose deficits.

Proffered Papers: RTT 3: Impact of variations on treatment planning

OC-0303 Dosimetric benefit of a clinically applied adaptive plan selection strategy for rectal cancer R. De Jong 1 , J. Visser 1 , N. Van Wieringen 1 , K. Crama 1 , J. Wiersma 1 , D. Geijsen 1 , A. Bel 1 1 Amsterdam UMC, Radiation Oncology, Amsterdam, The Netherlands Purpose or Objective For rectal cancer radiotherapy, an adaptive strategy by means of plan selection was clinically implemented in May 2016. We evaluated target coverage and dose to the organs at risk for the clinically used adaptive plan selection strategy compared to a non-adaptive approach for both short (5x5Gy) and long (25x2Gy) treatment schedules for rectal cancer patients. Material and Methods For this study the first 20 consecutive patients treated between May and September 2016 were included, if the length of the upper part of the mesorectum, as measured from the base of the bladder, was over 4.5 cm. For each patient 3 plans were created with different ventral PTV margins to the upper mesorectum ( fig 1 ), i.e. the most mobile part of the target volume. The chosen margins depended on status of rectal filling on the CT scan: 25/15/0 mm for empty rectum versus 15/0/-15 mm for full rectum. All patients were planned with VMAT. Based on daily Conebeam CT (CBCT) scans RTTs selected the plan with the smallest PTV that encompassed the complete target volume for treatment. These plans were compared to a non-adaptive strategy with a single plan and a ventral PTV margin of 20 mm. For each fraction bowel cavity, bladder and target volume (mesorectum) were delineated on the CBCT scan by a single observer. The dose

Fig 2 shows the evaluation of ΔD95% during simulated treatments with and without dose-guided couch corrections. The mean (range) of the final ΔD95% was 5.9 %-point (1.0-26.2 %-point) without inter-field couch corrections and 3.8 %-point (1.0-10.0 %-point), and 3.1 %- point (0.6-9.5 %-point) with dose-guided couch corrections using 10% and 5% ΔD95% thresholds, respectively. The mean number of dose-guided couch shifts per fraction was 0.4 (10% threshold) and 1.1 (5%).