ESTRO 36 Abstract Book

S258 ESTRO 36 2017 _______________________________________________________________________________________________

OC-0490 A robust and fast planning approach for adaptive MR-guided treatment of pancreatic cancer O. Bohoudi 1 , A. Bruynzeel 1 , B. Slotman 1 , S. Senan 1 , F. Lagerwaard 1 , M.A. Palacios 1 1 VUMC, Radiotherapy, Amsterdam, The Netherlands Purpose or Objective In May 2016, we implemented stereotactic MR-guided adaptive radiation therapy (SMART) using the MRIdian system (Viewray) for locally advanced pancreatic cancer. Interfractional changes in the anatomy of adjacent organs- at-risk (OARs) make daily online plan adaptation desirable. The main challenge of online plan adaptation is the requirement that it must be performed fast while the patient remains in treatment position. We evaluated an in-house developed re-planning strategy, which is Before use of SMART, robust baseline IMRT plans for online re-optimization are first produced with the MRIdian planning system (ViewRay). The same planning software is available at the treatment console for plan adaption. The target structure used for optimization is defined as PTV opt (GTV+3mm minus OARs). OAR contours are then spatially partitioned in separate OAR portions located within 1, 2 and 3cm from the PTV OPT surface, thereby allowing direct control over the spatial dose distribution (Fig. 1). The optimization process relies on a model which predicts OAR dose as a function of distance from PTV OPT , and generates optimization objectives to achieve a robust baseline plan for daily adaption. For daily SMART, physicians only re- contour OAR located within 3cm from PTV opt , based on the daily MR imaging (SMART 3CM ). Optimization structures are automatically adapted to the new anatomy, and re- optimization is performed using exactly the same plan parameters. This limited re-contouring strategy was evaluated by comparing 45 previously delivered fractions against a simulated standard (re-)planning method using full-scale OAR (re-)contouring, where optimization objectives were used for the whole organs (SMART FULLOAR ). Baseline plans for OAR were created that had identical plan quality as were achieved for SMART 3CM . Efficiency of both strategies was scored according to the number of optimizations needed to generate a high quality plan. Plan quality was assessed using PTV coverage (V95%) and institutional OAR constraints (V33Gy and V25Gy). currently in clinical use. Material and Methods

fractions (median V95%=93±6.8%). SMART 3CM uniformly resulted in plans which complied with the V33Gy dose constraint for OARs, whereas SMART FULLOAR failed in 35% of the cases to adhere to the V33Gy dose constraint according to the clinical protocol. Both strategies achieved V25Gy values lower than 20 cc for all OARs in every fraction. However, on average, SMART 3CM resulted in a lower V25Gy than SMART FULLOAR (Fig.2).

Conclusion This fast and robust (re-)planning approach for SBRT to pancreatic tumors requires clinicians to only re-contour OARs located within 3cm of the PTV OPT. Spatially partitioned optimization structures within this 3 cm region allowed for optimal OAR sparing, and adequate target coverage, using exactly the same plan parameters. OC-0491 Quality assurance of a novel table mounted imaging device integrated in a patient positioning system A. Utz 1 , A. Ableitinger 1 , A. Zechner 1 , M. Mumot 1 , M. Teichmeister 1 , P. Steininger 2 , H. Deutschmann 2 , M. Stock 1 1 EBG MedAustron GmbH, Medical Physics, Wiener Neustadt, Austria 2 medPhoton GmbH, Medical Physics, Salzburg, Austria Purpose or Objective Image guided radiation therapy (IGRT) aims to reduce margins and subsequently increase dose sparing for OAR. The majority of image guidance procedures are based on ceiling/floor- or gantry mounted imaging devices. In our particle therapy center a novel approach for patient alignment was introduced. The imaging system (imaging ring) is mounted on the treatment table and as such, allows high imaging flexibility e.g. CBCT or planar imaging at different table positions. The goal was to establish a phantom and a concept for a quality assurance procedure for the whole IGRT workflow. Material and Methods The IGRT Phantom consists of a PMMA cube with steel fiducials, which can be placed in predefined offset positions on a baseplate to simulate clinical patient shifts. An additional support structure is used to lift the cube. Holes on the upper corners of the cube allow to independently determine the absolute position with a lasertracker (see figure 1a). A CT imageset of the cube in a reference position serves as planning CT. The baseplate

Results The SMART 3CM

baseline plans required a lower number of

optimizations than SMART FULLOAR

(4 vs 17 on average).

PTV OPT

coverage with both strategies was identical in all