ESTRO 38 Abstract book

S1002 ESTRO 38

dysphagia, and no other symptom. Step 6: follow-up after PT. Three months after PT, the patient is PS 0, with no symptom, normal blood tests and metabolic CR according to the LUNAGO criteria on PET-scan (DEAUVILLE score at 2). Conclusion The use of PT for HL engaging the mediastinum is promising and there is a dosimetric advantage of reducing the dose to OAR. However, the limited availability of PT calls for case selection based on a clear understanding of which patient will most benefit from PT compared to advanced photon techniques. EP-1846 Cloud-based contouring education system supporting access from multi-devices H. Takegawa 1,2 , S. Nakamura 1,2 , N. Tanigawa 1 1 Kansai Medical University, Department of Radiology, Osaka, Japan ; 2 Kansai Medical University Hospital, Division of Radiation Oncology, Osaka, Japan Purpose or Objective Contouring uncertainties is sensitive to clinical outcomes in intensity-modulated radiation therapy due to its steep dose gradient around tumors and organs at risk. To reduce contour uncertainties among radiation oncologists and institutions, contouring seminars and workshops are held. However, contouring education systems used in those seminars does not often support access from multi- devices and multi-platform environments. The purpose of this study was to develop a cloud-based contouring education system supporting access from multi-devices. Material and Methods The system overview is shown in figure 1. A fundamental desktop application of the system was developed in MATLAB language and basic functions of contouring education were implemented as described below. It is considered that the system is used by experts(teachers) and beginners(students). Experts prepare educational data with their treatment planning systems (TPSs) at their institution and send them to the cloud system and then beginners can access these data according to their privilege. Educational image resources such as treatment planning CT, MR and PET/CT are imported, and image fusion are performed based on a DICOM image registration object exported from a TPS to ensure same image registration results as those of TPS. Data of structure sets which experts prepared are also imported to the system. Beginners draw contours on planning images without any contours in the structure sets and experts can modify beginners' contours to give them feedback. Educational effects can be measured by evaluating dice similarity coefficients of contours between experts and beginners. To allow users access the developed desktop application from multi-devices, it was deployed to a cloud system, Amazon Workspaces, which is a virtual desktop infrastructure (VDI) on Amazon Web Services. Access from desktop and laptop computers, and tablet with the VDI client application and from HTML5- compatible web browsers to the system were verified.

Figure 1. Overview of the cloud-based contouring education system. Results All devices successfully accessed the system with the VDI client and the deployed application worked without any problems. A modern input device such a stylus pen also worked and so it was possible to contour targets and organs at risk as drawing pictures. The delay in display and input were quite less even though the access was performed through 4G mobile network. Not all the web browsers accessed the system and input with a touch screen did not work. Conclusion The cloud-based contouring education system supporting access from multi-devices was developed and its access from multi-devices was verified. EP-1847 Automatic optimization of Head and Neck treatments with multicriteria radiobiological cost functions M. Esposito 1 , A. Ghirelli 1 , S. Pini 2 , S. Russo 1 , G. Zatelli 3 , P. Alpi 4 , R. Barca 4 , S. Fondelli 4 , B. Grilli Leonulli 4 , L. Paoletti 4 , F. Rossi 4 , P. Bastani 4 1 Azienda Sanitaria USL Centro, S. C. FIsica Sanitaria- FIrenze Empoli, Bagno a Ripoli, Italy ; 2 Azienda Sanitaria USL Centro, S. C. Fisica Sanitaria-Firenze Empoli, Bagno a Ripoli, Italy ; 3 Azienda Sanitaria USL Centro, S.C. Fisica Sanitaria-Firenze Empoli, Bagno a Ripoli, Italy ; 4 Azienda Sanitaria USL Centro, S. C. Radioterapia, Bagno a Ripoli, Italy Purpose or Objective The optimization of Head and Neck treatments is one of the most complex task for planners; skills and experience can influence planning results. Automatic planning can help standardize plan quality and is needed for adaptive radiotherapy. Aim of this study is to construct and evaluate a template for automatic optimization of Head and Neck treatments using multicriteria radiobiological cost function of the treatment planning system Monaco 5.10. Material and Methods The automatic workflow was tested on ten previously treated Oropharyngeal Head and Neck cases. Each treatment consisted in multiple prescription target volumes (PTV): high dose PTV (between 66 and 70 Gy), intermediate dose PTV of 60 Gy and a low dose PTV of 54 Gy in 32 fractions. Constraints for organ at risk (OAR) were: Parotid D mean < 26 Gy, Cord planning risk volume (PRV) D 1cm3 <50 Gy, esophagus D mean < 34 Gy, the patient D max <110% of the higher PTV dose. Planning objectives were to cover at least the 95% of each PTV with the 95% of the prescription dose. The radiobiological functions used in the template are shown in Table1. To each function a dosimetric objective was assigned; when all