ESTRO 38 Abstract book

S63 ESTRO 38

by the many customised devices designed and produced for implant optimisation over the years. In this context 3D printing is just a new tool for individualising the implant template in accordance with the distinctive patient anatomy and in close proximity to the clinical workflow, needs and possibilities within the individual departments. Several materials for 3D printing have the advantage of being MRI compatible, biocompatible, autoclavable and sufficiently cheap to allow for single use. However, in relation to safety and legal aspect it should be emphasized that the 3D printed product never should carry the radioactive source directly and only should serve as a platform for optimization of insertion and angulation of approved needles. Whereas commercially available templates can cover many clinical situations well with a standard implant configuration, 3D printed templates for steering of needles are in particular useful when individualised needles tracks are required. However, in such cases the local anatomical conditions should provide a reproducible and robust anchoring of the 3D printed template both with regard to rotation and angulation such that a meaningful pre-plan with virtual angulated needle tracks can be made. Ideally such a pre- plan should be based on MRI with a template in situ performed in a separate session not used for treatment. There is currently very limited data on the clinical use of 3D printed templates for steering of needles. This lecture will seek to review this experience using locally advanced cervical cancer as a practical example of the involved workflow, patient selection and possible outcome. Symposium: Automatic / Knowledge based treatment planning:open issues SP-0128 How can we get the best out of knowledge- based planning? S. Currie 1 , E. Miguel-Chumacero 1 , N. Laverick 1 , G. Currie 1 1 NHS Greater Glasgow And Clyde, Rt Physics, Glasgow, United Kingdom Abstract text Knowledge Based Planning (KBP) is now a recognised tool in the creation of high quality and consistent treatment plans. KBP was introduced as the standard planning approach for the majority of radical clinical sites at the Beatson West of Scotland Cancer Centre over three years ago based on the Varian RapidPlan solution. The presentation will describe the process of model implementation including the various barriers and pitfalls that were experienced. Additionally, the challenges faced in the continuing management of KBP models which now underpin over 50% of the daily clinical planning workload, across new software releases and updates, will be considered and how risks can be managed and addressed. Following the implementation of KBP, a new development in treatment plan optimisation became available through Multi-Criteria Optimisation. This allowed a new strategy to be adopted in the creation of knowledge based planning models by combining Varian’s Multi-Criteria Optimization- based Trade-Off Exploration to create base plans of the highest plan quality. These plans were optimised with emphasis in producing plans that offered reduced Organ at Risk (OAR) doses while always maintaining adequate and acceptable coverage of the planned tumour volume. These higher quality plans were then selected for the generation of new and updated knowledge based planning models. Results will be presented on the comparison of KBP plans generated with and without Trade-Offs, describing dose-volume parameters for both organs at risk and the PTVs, and target homogeneity, conformity, and plan complexity and deliverability. Finally, the opportunity afforded by knowledge based planning tools to aid in decision support will be

Symposium: Real time navigation technologies in brachytherapy

SP-0125 Multi-modal Image Fusion to support Minimally-invasive Therapy D. Barratt 1 1 University College London Ucl, Department Of Medical Physics and Biomedical Engineering / Ucl Centre For Medical Image Computing / Wellcome Trust EPSRC Centre for Interventional and Surgical Science, London, United Kingdom Abstract text With the increasing availability of new treatment approaches for prostate cancer based on minimally- invasive surgical and radiation therapies, there is a growing need for computer-assisted therapy systems that enable high-accuracy, tumour-targeted therapy delivery within an easy-to-use and efficient clinical workflow. In this presentation, challenges to high-accuracy therapy delivery and clinical usability using computer-assisted planning and guidance systems will be discussed, with a focus on the impact and compensation of soft-tissue motion (including physical deformation). Approaches to automatic generation and application of computer- generated deformable organ models to compensate for organ motion within image fusion software will be described, drawing on the presenter's prior research in this area. Methodological aspects of validation and, in particular, accuracy evaluation, for such systems will then be discussed. Finally, the recent and potential future impact of contemporary machine learning techniques, such as convolutional neural networks, will be explored. Abstract text It is challenging to steer needles accurately too difficult to reach places. In the first part of the presentation a short overview will be given of different principles to steer needles. Examples from hand held steerable needles and robotically steerable needles presented. Second part of the talk will focus on a novel way to steer needles during brachytherapy treatment of cervical cancer, in cases where the tumour is extending lateral to the pelvic wall or away from the vaginal surface. In these situations the use of an applicator is not sufficient to cover the target area, and free hand steering of needles is difficult and may result in incorrect placement or steering for a long distance through healthy tissue. To overcome this, a method will presented based on personalised 3D-printed vaginal templates with curved channels. These curved channels are then used to steer needles. The shape of the template will be based on an MRI scans. In this study a template was printed having channels with different curvatures to examine the effect of curvature constraints on placement accuracy. The feasibility of needle steering using curved channels in a template was tested in an experimental set-up using gelatin phantoms. SP-0127 3D printed templates for steering of needles J.C. Lindegaard 1 , P. Petric 1 , A. Traberg Hansen 1 , S. Kynde Nielsen 1 , B. Meisner 1 , K. Tanderup 1 , L.U. Fokdal 1 1 Aarhus University Hospital, Department Of Oncology, Aarhus C, Denmark Abstract text Standard templates for needles guidance have been used successfully for many years in brachytherapy. Brachytherapy has always been a field of innovative activity towards personalisation as exemplified SP-0126 Steering of needles and applicators J. Dankelman 1 1 Delft University of Technology, Biomechanical Engineering, Delft, The Netherlands