Abstract Book

S363

ESTRO 37

pain response compared to 53% and 60% for Conv-8Gy and DP-16Gy. QoL analysis suggests better outcome for the DP-8Gy arm with the scores ‘painful characteristic’, ‘insomnia’ and ‘appetite loss’ reaching significance (p<0.05). DP-8Gy was selected for phase-III investigation. Despite a 30-year old concept [1], a solid rationale [2] and availability of methods for DP since over a decade [3- 5], a limited amount clinical data is available. DP offered a small window for dose escalation to radioresistant regions in head and neck cancer at the expense of increased dysphagia and mucosal injury [7, 11]. Longitudinal per-treatment imaging and treatment adaptations result in heavy workloads and high cost rendering the future of this approach unclear. Randomized clinical trials are under way [10, 12-14]. Other applications using single imaging in single or oligo- fractionated RT are less costly and more promising. DP for reducing organ toxicity has hardly been explored although strong physics and biological rationales exist [15-17]. 1.ActaOncol.1987;26(5):377-85. 2.IJROBP.2000; 47:551- 560. 3.PhysMedBiol.2003;48(2):N31-5. 4. PhysMedBiol.2006;51(16):N277-86. 5. Radiother Oncol.2006;79(3):249-58. 6. Head Neck.2017;39(11):2264-2275. 7. ActaOncol.2017:8:1-7 . 8.Cancer.1985;56(5):979-90. 9. R&O.2017;122(1):30-36. 10. J.Med.Im.Radiat Oncol.2017;61(1):124-132. 11. R&O.2016;120(1):76-80. 12. Trials 2011;12:255. 13. R&O 2012;104:67-71. 14. BMC Cancer 2013;13: 84. 15. Cancer.2016;122(13):1974-86. 16. Radiat Oncol.2015:30;10:72. 17. MedPhys.2017;44(7):3418-3429. SP-0694 For the motion Autoplanning will present a new life for RTTs! I. Kristensen 1 1 Skåne University Hospital- Lund, Radiation Physics, Lund, Sweden Abstract text Auto-planning is an extension and development of knowledge-based planning. Good quality treatment plans are required and have to be available to the model auto- planning is based on. There are several important advantages with auto- planning which should be considered; *Uniformity – auto-planning will produce plans with less variation among dosimetrist and between them patient resulting in a uniform plan quality *Will allow staff to spend more time on difficult cases/plans, delineation, research and development, educating colleagues *Better sparing of organ at risk For this to happen, we will have to produce (select) a number of treatment plan of very good quality, of a variety of patients (anatomy wise). This is needed, thus the generated treatment plans will be able to reflect the local technique / tradition / standards / dose-volume constraints. A larger library will aid the auto-plan model to produce good quality treatment plans. This is where we, the RTT’s experienced in treatment planning, comes in; *We will have to select/create those very good clinically approved quality plans-patients combinations for the models library. *We will evaluate the auto-plans; To compare the auto- generated plans to those manually optimised *And when decided in the institute that auto-planning will be used; To evaluate auto-plan before delivery to the patient Debate: Autoplanning, is there still a bright future for RTTs after automation?

*We will be a part of the development of new areas for auto-planning and the safe implementation of these new areas. *We will also end up with more time to manually optimise the more difficult cases. Auto-planning will put demands on the radiation oncologists and medical physicists as well. Dose-volume constraints for organs-at-risk, very specific ones including the prioritizing order between them will be needed. Decisions on the conformity of the target, as well as a priority-list for different areas will have to be made. What trade-offs will be accepted? And the auto- generated plans will have to be checked and maybe even adjusted to fit the local plan quality “rules”. Here the experienced RTT’s will have a role. Going back in time a bit, we started out with hand calculations. A slow process, thus we were not able to create treatment plans for all patients, for many patients, a mid-point calculation would have to do. Late seventies and early eighties we, with the introduction of computers, started to see both in-house built and commercial treatment planning systems. The need for RTT’s in treatment planning would be smaller, the computers could do it all, they said. For the benefit of the patients; we started to create treatment plans for more patient groups, and even more complex treatment plans. We also saw a rise in the number of patients in our departments. So we needed more RTTs for treatment planning. When we first started to talk about inverse planning – a press on a button would do it – no need for RTT’s. Well, this is not what happened. IMRT turned out to be quite complex to plan. The need for qualified and experienced staff creating treatment plans rose. VMAT planning can be quite complex as well. And will not be done with the press of a single button. There will be plenty for treatment planning RTTs to see too in the future! SP-0695 Against the motion L. Johnsen 1 1 Odense University Hospital, Radiofysisk Laboratorium, Odense, Denmark Autoplan Through the history of radiotherapy, both planning and treatment complexity, has increased in order to reach better tumour control and decreased toxicity. Since the 80’s, the treatment planning process has seen a tremendous increase in use, mainly due to computational power. In the later years there has been an increase of automation of many processes throughout the RT workflow, increasing quality and reducing time consumption. Several vendors of treatment planning systems (TPS) are now offering automatic planning (AP) software, that optimize dose distribution without human input after initialization. The algorithm is able to search a larger solution space than a human dose planner and thereby able to consistently produce plans of equal or higher quality than experienced RTT’s (and physicists and dosimetrists). This has already been shown in multiple publications for several indications such as head and neck, esophagus, prostate, etc. Although this does not hold for all indications yet, AP is still a new technique in RT and under intense development. To predict the role of the RTT’s in dose planning, we need only look at other professions where automation has taken over the workload: In the car industry, for example, robots are building cars faster and with higher precision and at a lower cost than humans did. As a result jobs have been cut away. And we may as well prepare for the same scenario for RTT’s in the coming years. Initially, we are seeing AP and manual planning being combined, but as confidence in AP increases an increased number of processes will be allowed in clinical practice. With plan Abstract text Rtts role after automation and