ESTRO 38 Abstract book

S1108 ESTRO 38

enable accurate and efficient adaptive radiotherapy for head and neck cancer patients. Material and Methods Material/methods: Patients treated on an IRB approved adaptive radiotherapy trial using weekly CT and MR and daily CBCT guidance were evaluated. The accuracy of a commercial deformable registration algorithm, using correlation coefficient and weighted Dirichlet energy, was evaluated for CT to CT and CT to CBCT DIR. For CT to CT DIR, 9 patients with mid-treatment repeat CT scans were obtained. Normal tissues were contoured on both images. The contours were propagated using DIR from the planning CT onto the mid-treatment CT and compared to the clinician drawn contours using the Dice similarity coefficient (DSC). The performance of the CT to CBCT DIR was evaluated relative to the CT to CT DIR, by rigidly registering the CBCT obtained the same day as the mid- treatment CT and copying the contours onto the CBCT. DIR was performed between the planning CT and the mid- treatment CBCT and the propagated contours were compared using DSC. The accuracy of dose calculation on the CBCT was evaluated by comparing the difference in the clinically relevant metrics between the dose calculated on the CT and the CBCT obtained on the same day. Dose accumulation was performed based on only weekly CTs and based on daily CBCTs. Differences in total accumulated dose over treatment was evaluated. Results Results: CT to CT DIR accuracy was evaluated for the brainstem, larynx, parotids submandibular glands, mandible and esophagus using DSC. The average DSC was 0.80 (SD 0.05), range: 0.73 (esophagus) to 0.91 (mandible). CT to CBCT registration accuracy was evaluated for 10 image pairs. The average DSC, relative to the CT to CT DIR, was 0.91 (SD 0.04), range: 0.89-0.94. Dose difference using CT or CBCT was evaluated for 10 image pairs based on the GTV, CTV, submandibular glands, larynx, mandible, and parotids. The average difference of clinical metrics was 0.6 cGy (SD 2.7 cGy) or 0.07% of the planned fraction dose (SD 2.7%). The maximum difference was 8 cGy, observed for a single image pair for the mean larynx dose. Based on daily CBCT dose accumulation, the average difference between the planned dose and the delivered dose for the targets was -170 cGy (SD: 51 , range: -260 - -97) and for the normal tissues was -51 cGy (SD: 143 cGy, range: -277 – 181 cGy). For the target structures, the average difference between the dose accumulated using daily CBCT compared to weekly CT was 114 cGy (SD: 39 cGy, range: 55 - 180 cGy). For the normal tissues, the average difference between the dose accumulated using daily CBCT compared to weekly CT was 53 cGy (SD: 93 cGy, range: -101 – 175 cGy). Conclusion Conclusion: Dose accumulation is clinically feasible and can be used to guide adaptation to ensure target coverage and normal tissue sparing and to improve understanding of delivered dose and outcomes. EP-2022 Dose-dependent changes in T2w-MRI texture of obturator muscles after prostate cancer radiotherapy E. Scalco 1 , T. Rancati 2 , A. Mastropietro 1 , A. Cicchetti 2 , B. Avuzzi 3 , R. Valdagni 2,3,4 , G. Rizzo 1 1 Istituto di Bioimmagini e Fisiologia Molecolare, CNR, Segrate Milano, Italy ; 2 Fondazione IRCCS Istituto Nazionale dei Tumori, Prostate Cancer Program, Milano, Italy ; 3 Fondazione IRCCS Istituto Nazionale dei Tumori, Radiation Oncology 1, Milano, Italy ; 4 Università degli Studi di Milano, Department of Oncology and Hemato- oncology, Milano, Italy Electronic Poster: Physics track: Quantitative functional and biological imaging

based on planar kV-MV, in-room lasers or surface scans. The different setup techniques were evaluated using cone beam computed tomography (CBCT) as the ground truth. Material and Methods This prospective clinical trial included 39 patients, from which 102 treatment fractions were analysed. Patients were treated with deep inspiration breath-hold radiotherapy after breast conserving surgery, without lymph node involvement. The initial setup was based on in-room lasers and tattoo marks. A surface scan was acquired (Catalyst and Sentinel, from C-RAD Positioning AB, Uppsala, Sweden) followed by a CBCT scan. Lastly, the patient was moved to the planned treatment position based on anterior-posterior kV and tangential MV images (kV-MV), and treatment was delivered. In off-line analysis, the planned treatment position was found based on the acquired surface scans and on the acquired CBCTs scans (soft tissue, semi-automatic rigid registration in six degrees of freedom). The treatment position found from the CBCT scans were used as the ground truth. The CTV to PTV margins were calculated according to van Herk [1] using the systematic (Σ) and random (σ) errors between the different setup techniques (kV-MV, in-room lasers and surface scans) and the CBCT. Results CTV to PTV margin required to ensure dose coverage of the CTV were largest in the vertical direction for all setup techniques (Table 1). Overall the margins were smaller for patient positioning based on kV-MV compared to in-room lasers and surface scans. The margins were smaller in the lateral and vertical direction using surface scans compared to in-room lasers. Of the 102 treatment fractions, the number of treatment fractions with a setup error above 1 cm in either direction was 6, 4 and 1 for in- room lasers, surface scans and kV-MV, respectively.

Conclusion The study shows that the CTV to PTV margin is smallest for a kV-MV setup and suggests that a CTV to PTV margin of 6, 7 and 9 mm should be applied in the lateral, longitudinal and vertical directions, respectively. The margins from the initial setup based on in-room lasers can be reduced in the lateral and vertical direction using surface scans; in the longitudinal direction the margins are similar. [1] van Herk M, Remeijer P, Rasch C, et al. The probability of correct target dosage: Dose-population histograms for deriving treatment margins in radiotherapy. Int J Radiat Oncol Biol Phys 2000;47:1121–1135. EP-2021 Commissioning and clinical implementation of dose accumulation and adaptive radiotherapy K. Brock 1 , M. McCulloch 1 , G. Cazoulat 1 , A. Ohrt 1 , P. Balter 1 , H. Bahig 1 , S. Ping 1 , A. Mohamed 1 , H. Elhalawani 1 , B. Elgohari 1 , S. Frank 1 , J. Wang 1 , D. Rosenthal 1 , C. Fuller 1 1 The University of Texas MD Anderson Cancer Center, Imaging Physics, Houston, USA Purpose or Objective Purpose/Objective: To commission a commercial deformable registration and dose accumulation system to