S90 ICTR-PHE – 2014 phantom is in static mode to achieve clinically acceptable gamma pass rate values (> 95%) for head-and-neck and prostate with pelvic lymph node RapidArc delivery verification using 6, 10 and 15 MV photons. Keywords : RapidArc, Octavius 4D, Rotational Dosimetry 183 The use of the Golden Triangle paradigm for Knowledge Exchange for Computational Radiotherapy: a Case Study M. Simmons 1 , N. Burnet 2 , A. Parker 3 1 Cavendish Laboratory, University of Cambridge, UK 2 Department of Oncology, University of Cambridge, UK 3 High Energy Physics Group, Cavendish Laboratory, University of Cambridge, UK Purpose: Cern initiatives in developing medical applications for its technology have been significant drivers enabling us to develop collaborations in this field, where experts from particle physics, medical physics, computer science, engineering, and medicine are working together to make advances in radiation oncology. Such interdisciplinary teams, we argue, are able to spark innovation in ways which otherwise would be impossible. We present the development of collaborative Computational Radiotherapy initiatives as a Case Study. These collaborations at Cambridge and beyond are examples of “[those] most exciting collaborations [which] arise as a result of like-minded people getting together – sometimes by chance – to address a problem.” (From the UK Treasury Lambert Review of 2003.) Methods: The talk presents the powerful and novel Golden Triangle paradigm for Knowledge Exchange (KE) – encompassing people, ideas and funding. Our application of this paradigm has underpinned the instigation and growth of Computational Radiotherapy, and our concomitant strategies for facilitating effective knowledge exchange. The creation and maintenance of a web of effective relationships has been indispensible. Scientists from different disciplines do not routinely meet nor speak the same ‘language’, any successful application of knowledge exchange must address this and break down silos. We make a persuasive case that KE is overwhelmingly a human-centric art rather than a data- centric science. We present evidence which suggests that undue reliance and interfacing with data systems is positively detrimental to effective knowledge exchange. Examples cited will include the effectiveness, or otherwise of email. Results: The talk looks at progress so far. It will also discuss our mutually beneficial links with companies and other organisations. The talk covers the highlights of the collaborations built on the human web, as well as some of the lessons learnt. These include the contribution of Cern techniques in warehousing and managing extremely large volumes of data, critical in our winning the bid for our flagship research project, VoxTox. and an essential elements in our other two projects, Accel-RT and GHOST. We assess the overall impact of our collaborations in radiation oncology. Conclusions, and the future: Finally, we scan the horizon for future possibilities in computational radiotherapy. Additional challenges include public education and information about the societal value of physics research and radiation oncology research. Also, how can we build on our experience of human network building so far to make them more effective? We have learnt that knowledge cannot simply be encoded as digitised data, but must be communicated between humans if any sustainable progress is to be made. Good human-centric facilitation is essential for effective knowledge exchange. Keywords: people-centric, data-centric, collaborations 184 Prompt gamma imaging of proton pencil beams at clinical beam current J. Smeets 1 , A. Celani 2 , E. Clementel 3 , C. Fiorini 4,5 , T. Frizzi 2 , S. Henrotin 1 , G. Janssens 1 , I. Perali 4,5 , D. Prieels 1 , F. Roellinghoff 1 , F. Stichelbaut 1 1 Ion Beam Applications SA, Louvain-la-Neuve, Belgium 2 XGLab, Milano, Italy 3 iMagX Project, ICTEAM Institute, Université Catholique de Louvain, Louvain-la-Neuve, Belgium 4 Politecnico di Milano, Dipartimento di Elettronica, Informazione e Bioingegneria, Milano, Italy 5 INFN, Sezione di Milano, Milano, Italy Thanks to its concentrated dose deposition pattern, proton therapy offers precision and consequently demands accuracy. In order to further improve feedback on treatment delivery, different concepts of cameras are being designed in view of instantaneous range verification by imaging prompt gammas emitted along the proton tracks in the patient. Prompt gamma imaging is a simple concept applied in most challenging conditions. Technical requirements include (1) large dynamic range for the broad, continuous, multi-MeV energy spectrum of prompt gammas, (2) high detection efficiency to take most benefit of the few prompt gammas emitted per individual proton pencil beam spot, (3) low sensitivity to the uncorrelated neutron background and (4) very high count rate to score several thousand counts in the typical millisecond duration of a beam spot. The present study reports on the development and test of a prompt gamma camera relying on a slit collimator to produce a reversed 1-dimensional projection of the last centimeters of the beam path on a segmented LYSO detector for treatments delivered in pencil beam scanning mode. The geometry was optimized by means of simulations to reach millimeter accuracy on range estimation for doses compatible with single pencil beams by comparison of a measured profile with a reference simulated one. In order to meet the demanding specification on the count rate capability, a dedicated, cost-effective photodetection system was designed. This very-fast, 1-dimensional, high- energy gamma imaging device relies on two rows of 20 LYSO crystal slabs, directly coupled to SiPMs’ arrays and readout by 40 independent acquisition channels in fast counting mode. A first prototype limited to 3 channels was implemented and validated in April 2013 during proton irradiations of a PMMA target at the West German Proton Therapy Centre in Essen. This prototype was, to our knowledge, the very first to achieve successful acquisitions of correlated prompt gamma profiles at clinical beam currents of several nA at nozzle exit. These performances were reached at close distances and up to the maximum clinical beam energy of 230 MeV, demonstrating the technical feasibility of prompt gamma imaging with both sufficient sensitivity and counting speed for application in pencil beam scanning mode. Leveraging this proof of concept, the full-size photodetection system equipped with the first of its two rows of 20 slabs was assembled and tested in October 2013. Correlated profiles were recorded for individual pencil beams of a few milliseconds duration. A count rate of 1 MHz per individual slab was even reached above a 2 MeV threshold for a beam