ESTRO 2021 Abstract Book

S373

ESTRO 2021

In order to start building a federated radiomics infrastructure, we previously performed a proof-of-concept study Shi and Zhovannik et al where two Dutch centers’ data was used to repeat the study of Aerts et al . This study used only four original radiomic features from Aerts et al article. The Resource Description Framework (RDF) triple database was used to store clinical and imaging data locally according to Radiation Oncology Ontology (ROO) and Radiomics Ontology (RO). In the current project, one more center joined the original federated radiomics network. 100+ radiomic features and the clinical outcome were mapped by researchers locally in accordance with ROO and RO. We developed federated machine learning algorithms for the federated radiomics network, the algorithms are able to be trained without exposing the sensitive imaging data preserving privacy.

SP-0486 SBRT in clinical practice for pancreatic tumours M. Kountouri 1 1 Geneva University Hospitals, Radiation Oncology, Geneva, Switzerland

Abstract Text Pancreatic adenocarcinoma remains a lethal malignancy, being the gastrointestinal tumour with the most sombre prognosis and the fourth most common cause of cancer mortality in Europe. Surgery is considered the only possible curative treatment option for these patients although less than 10% of patients are operable at diagnosis. The role of radiotherapy remains controversial, often used in combination with chemotherapy in the consolidation of the treatment in locally advanced pancreatic cancers, resulting in an improved progression free survival (PFS). Refined radiotherapy treatment techniques, aiming to improve target definition and motion management, allowed for an increase in the radiotherapy dose per fraction while limiting the toxicity to the surrounding tissues. Stereotactic body radiotherapy (SBRT) nowadays is considered a treatment option in patients with locally advanced pancreatic cancer who do not progress following systemic therapy. Early phase II trials using extreme hypofractionation schedules demonstrated an improved local control at the expense of substantial gastrointestinal toxicity associated with reduced overall survival (OS). As a result, further studies of fractionated SBRT with lower single fraction dose have been proved to be less toxic and, when integrated with chemotherapy, to lead even in improved OS. De Geus et al (Cancer, 2017) in one of the largest retrospective studies, including 14,331 patients from the National Cancer Data Base, showed that SBRT following chemotherapy in unresectable pancreatic adenocarcinomas improved OS compared to chemotherapy alone or compared to the addition of external beam radiotherapy or intensity modulated radiotherapy to chemotherapy. The next question that needs to be addressed is at what dose can we achieve the maximum disease control with minimum toxicity. There is still no consensus on this as there are no randomised trials available. Among the published data there is great variability on the total dose prescribed, the dose per fraction as well as the prescription isodose. A frequently reported dose is 40 Gy in 5 fractions (BED 10 72 Gy, BED 3 147 Gy) while doses of 3 fractions (total dose 30 – 45 Gy) and 5 fractions (total dose 25 -45 Gy) have also been described. Another key element in the successful treatment of pancreatic cancers with SBRT is the management of the motion. Insertion of at least three fiducial markers usually endoscopically is often implemented while biliary stents are controversial surrogates for tumour position. A 4-dimensional computed tomography (4DCT) scan is mandatory for all patients while contrast enhanced end-expiratory breath hold scans are also recommended in order to define an internal target volume (ITV) and gastrointestinal structure planning organ-at-risk volume (PRV). There is also the option to treat during free breathing and in that case appropriate compensation such as gating, tracking, compression, or a combination is recommended and according to tumour or fiducial movement on respiration amplitude-reducing methods are appropriate. Despite the limited consensus on SBRT of pancreatic adenocarcinomas, it is widely agreed that this treatment should be offered in high volume centres in order to achieve the best possible outcomes. The Mobility / Technology Transfer Grant of ESTRO allowed me to undertake a visit at the Erasmus MC Cancer Institute to observe this treatment at all the stages; from simulation to planning and execution. This knowledge has proved extremely valuable to our expanding department in a University hospital that is also a tertiary centre for hepatobiliary cancers.

SP-0487 Skin effects after breast radiotherapy: Focus on biological features A. Cicchetti Italy

Abstract not available

SP-0488 Transitioning from 2D to IMRT in head and neck cancer J. Paragas Philippines

Abstract not available

Poster highlights: Poster Highlights 18: Radiotherapy treatment outcome

PH-0489 How can we predict permanent xerostomia after radiotherapy in patients with head and neck cancer? M. Bruvo 1 , A.M.L. Pedersen 2 , E. Samsøe 3,4 , C. Maare 3 , F. Mahmood 3,5,6 1 University College Copenhagen, Department of Technology, Bachelors Degree in Radiography, Copenhagen,