ESTRO 2020 Abstract book

S305 ESTRO 2020

ESTRO Mobility Grant I had the opportunity to spend two weeks in Rennes at LTSI for an exchange of knowledge and methods with the French colleagues. The general aims of the visit were: to deepen knowledge on this topic, learning and applying optimal pre-processing steps to different MRI images (T2 weighted, Diffusion-weighted imaging and apparent diffusion coefficient maps) and training on extraction and management of RFs. Moreover, the ultimate goal was to customize the radiomic pipeline developed by the French team for the Italian cohort of patients, consisting in a magnetic resonance images (MRI) database from 20 patients undergoing brachytherapy (BT) for cervix cancer and having an MRI before each brachytherapy fraction, and to develop methods for the longitudinal analysis of those data. This unique aspect of the Italian patient cohort allows investigation of RFs extracted from three anatomical regions with different characteristics: (a) muscles (i.e. tissue exhibiting non- uniformity in structure and not receiving dose, so acting as a good check for test-retest of RFs); (b) bladder filling (i.e. region exhibiting uniformity in structure and not receiving dose, so acting as a good check for test-retest of RFs) and (c) cervix tumor (i.e. tissue exhibiting non-uniformity in structure and receiving dose, so acting as a good check for effect of dose on RFs). All images (77 scans) were acquired with two different scanners but with the same protocol, and this allows evaluation of inter-scanner variability of RFs. During the two-week visit I had the possibility to discuss with experts about some possible pitfalls in radiomics workflow and RFs analysis and to use new open- source software for imaging processing, segmentation and registration. Moreover, I have improved my understanding and broaden the knowledge on harmonization strategies and dimensionality reduction methods. The preliminary results achieved on the small available population (BT Italian cohort) led to promising results and opens the possibility to a first a joint methodological publication. Moreover, future studies with bigger datasets might be considered. Throughout this visit we strengthen the cooperation between professionals involved with radiomics studies and consolidate the complementary collaboration between universities and research hospitals. The mobility grant experience was the possibility for a cultural/technical exchange between two clinical and research environments with both complementary experience and common points. SP-0517 Increasing the therapeutic index of radiation by combination with rucaparib in cervical cancer S. Saha 1 , R. Howarth 1 , A. Dubrovska 2 , K. Marchbank 3 , N. Curtin 1 1 newcastle University, Newcastle University Centre For Cancer, Newcastle Upon Tyne, United Kingdom ; 2 oncoray - National Center For Radiation Research In Oncology, Biomarkers For Individualized Radiotherapy, Dresden, Germany ; 3 Newcastle University, Complement Therapeutics Research Group And National Renal Complement Therapeutics Centre, Newcastle Upon Tyne, United Kingdom Abstract text Cervical cancer is the 4th most common cancer in women worldwide. Despite vaccination, developing countries in the Sub-Saharan Africa and South-Eastern Asia are mostly affected. In the UK, the peak age of incidence is 25-30 years with a 10-year survival rate 63%. For the locally advanced disease common treatment regimen is cisplatin ± radiotherapy. However, cisplatin induced kidney toxicity results in discontinuation of the treatment and therapy

failure. The stage IVA disease is inoperable due to large tumour volume and candidates for curative radiotherapy only. Further, within these giant tumours poor availability of oxygen in blood flow results hypoxia induced radio- resistance. New therapeutic strategy is therefore urgently needed. Previous studies have shown that the PARP inhibitors (PARPi): Rucaparib TM , Olaparib TM , nicotinamide and benzamides may increase tumour blood-flow and improve tumour oxygenation. Cisplatin-induced kidney toxicity is thought to be PARP-mediated with PARPi may potentially protect the kidneys. To provide pre-clinical data to justify a clinical trial the objectives of my project are to study: (1) the chemo-radiosensitising ability of Rucaparib TM and (2) the ameliorating effect of Rucaparib TM on cisplatin-induced kidney toxicity. If successful, this study will contribute new chemo-radiotherapeutic strategies to reduce cisplatin and (or) radiation doses in advanced stages of cervical cancer. SP-0518 Development and validation of a delta- radiomics response model for neoadjuvant radiotherapy of soft tissue sarcomas J. Peeken 1,2,3,4 , E. Chen 5 , D.S. Hippe 2 , K. Specht 6 , E. Kim 2 , N.A. Mayr 2 , M.J. Nyflot 2 , S.E. Combs 1,3,4 1 klinikum Rechts Der Isar- Technical University Of Munich, Department Of Radiation Oncology, Munich, Germany ; 2 university Of Washington, Department Of Radiation Oncology, Seattle, Usa ; 3 german Cancer Research Consortium Dktk, Partner Side Munich, Munich, Germany ; 4 helmholtz Zentrum Muenchen, Institute Off Radiation Medicine, Munich, Germany ; 5 university Of Washington, Department Of Pathology, Seattle, Usa ; 6 klinikum Rechts Der Isar- Technical University Of Munich, Department Of Pathology, Munich, Germany Abstract text Introduction: In soft tissue sarcoma (STS) patients that receive neoadjuvant radiotherapy (RT) pathological measures from the resected tumor such as the percentage of viable cells (%VC) do not have a comparable predictive power as observed in other malignant entities (Schaefer et al. IJROBP 2017). Regarding the unfavorable prognosis of high-grade STS, the selection of non-responding patients for additional therapy may help to personalize therapy regiments. The aim of this project is the development of a delta radiomics prediction model as a potential novel radiation response parameter. Material&Methods: In the course of a previous radiomics project, databases with clinical information and pre- therapeutic MRI scans were generated at the home institution at the Department of Radiation Oncology at the Technical University of Munich and at the host institution at the Department of Radiation Oncology in Seattle, USA. The scope of this work was to analyze the availability of MRI scans conducted after RT but before surgery and to prepare them for radiomic analysis. Results: Several exclusion criteria were defined including certain sarcoma histologies, treatment regiments, imaging artifacts and availability of MRI sequences. A large patient cohort with pre- and post-therapeutic imaging studies could be curated. Most patients had both MRI sequences available. Manual segmentation was successfully performed for all patients on both extracted MRI sequences. In contrast to the home institution, almost 60% of all patients received chemotherapy (CT) sequentially in addition to RT. We were able to analyze the post-CT imaging studies, too. The %VC could be determined for the majority of patients. For the remaining patients, the