ESTRO 36 Abstract Book

S893 ESTRO 36 _______________________________________________________________________________________________

anatomy match. 4. On the ML images, the entire breast with a ring structure of 5 mm is projected. For 2 out of 30 patients, the actual patient contour was projected outside the ring structure for several consecutive fractions, resulting in a new CT scan and treatment plan (adaptive RT). For 2.7% of the fractions, the residual deviation in lung wall was > 5mm. 5. CBCT images (if necessary in breath hold) are only acquired and analysed if indicated by the physicist based on deviations on one of the above images. Based on these results, an imaging protocol for breast cancer patients has been developed. Results of this protocol for (about 450) patients treated in the last half year will be presented. Conclusion Based on the above results, the imaging protocol for breast cancer patients is now as follows: cine images are acquired during the first treatment fraction, and if breathing motion is > 5 mm, further instructions are given to the patient and extra cine imaging is performed. Everyday, on-line setup verification using clips on orthogonal kV imaging is applied, as well as the use of ML verification images to 1) check and if necessary correct for residual lung wall deviations and 2) apply adaptive radiotherapy using an additional external contour ring structure. CBCT images are acquired only if required by the physicist based on deviations in one of the above images. EP-1646 Impact of interobserver variability and setup uncertainty on dose in organs-at-risk V. Prokic 1 , F. Röhner 2 1 Koblenz University of Applied Sciences, Faculty of Mathematics and Technology, Remagen, Germany 2 University Hospital Bonn, Department of Radiology, Bonn, Germany Purpose or Objective Accurate target and organs-at-risk (OARs) contouring and accurate and reproducible patient setup are crucial for success in radiotherapy, in particular when volumetric- modulated radiotherapy (VMAT) with high conformality and steep dose gradients is applied. Interobserver variability in contouring of OARs can have strong impact on dose-volume histograms, as well as possible setup- errors when setup-correction with on-board imaging is not performed daily. The goal of this study is to quantify the impact on delivered dose in parotid gland in patients with head-and-neck cancer irradiated in VMAT technic: (1) due to the interobserver variation in contouring of parotid glands and (2) due to the patient setup without daily image guidance. Material and Methods We have retrospectively analyzed seven patients who underwent primary definitive radiotherapy for head-and- neck cancer. Patient set-up is verified weekly using kV CBCT. The prescription dose was 50Gy/70Gy to PTV1/2. VMAT plans were generated using Eclipse 13.6 (Varian Medical Systems USA), for TrueBeam Linac with HD-120 MLC and 6MV. Plans were optimized to meet a set of dose constraints to OARs and the prescribed doses to the target volumes. Three radiation oncologist have independently delineated the parotid glands for this analysis. In order to estimate the impact of set-up errors on the dose in parotid glands that would have occurred without correction of patient positioning, for each patient the isocenter of the plan was shifted according to the weekly set-up error and the dose distribution was recalculated in treatment planning system for VMAT plan. Dosimetric impact due to the intraobserver and setup variations was quantified in terms of mean dose in parotid glands.

Figure1: Axial and coronal CT slices of FB (on the left side) and DIBH (on the right side) dose distributions for the same patient. Conclusion 96% of the patients trained were suitable for this technique. In 80% of the cases, DIBH treatment was chosen over FB. DIBH 3D RT for left breast cancer showed significant dose reduction for heart in every patient, and the left lung usually also benefits of it, achieving comparable coverage to PTV. EP-1645 Optimization of on-line setup verification and adaptive radiotherapy for breast cancer patients M. Essers 1 , S. Hol 2 , I. Maurits 2 , Y. Nijs 2 , T. Donkers 2 , L. Pontzen 2 , S. Toemen 2 , L. Mesch 2 , K. De Winter 2 1 Dr. Bernard Verbeeten Instituut, Department of Medical Physics, Tilburg, The Netherlands 2 Dr. Bernard Verbeeten Instituut, Radiothearpy, Tilburg, The Netherlands Purpose or Objective In recent years, new imaging modalities besides the long existing megavolt (MV) imaging using an electronic portal imaging device (EPID) have become available at the linac, to ensure accurate patient setup during the course of the radiotherapy treatment. The purpose of this work was to study the usefulness of all possible imaging modalities and to develop the optimal online imaging protocol for breast cancer patients in our institute. Material and Methods 1. kV or MV orthogonal imaging?: The interobserver variation in image registration was compared for 30 MV and 30 kV imaged patients. 2. Is cine acquisition useful, and how?: For 30 patients, cine images (“movieloop” during treatment fields) were acquired and analysed during all treatment fractions. 3. Surgical clips or bony anatomy for image registration?: For 30 patients, the setup changes based on patient anatomy and surgical clips registrations were compared. 4. How to use MV images in the mediolateral tangential direction (ML images) for adaptive radiotherapy? For 30 patients, the ML images during all treatment fractions were analysed . 5. Should we use CBCT images? Results 1. The interobserver variation reduced from 0.2mm for MV to 0.1mm (1 SD) for kV image registration. 2. Breath hold is very reproducible and stable within one fraction and during the course of fractions for the majority of patients . 1 of 30 patients was not able to have a stable (< 5mm) breath hold, for every fraction. 3. For 33% of fractions, the difference between anatomy and clip match was > 5mm, due to different patient (breast) posture or seroma or edema, or uncertainties in