ESTRO 38 Abstract book

S503 ESTRO 38

number of manually segmented brain structures associated with cognition were evaluated. Conformity relative to the target was assessed by using both the Conformity Index (CI) and Hausdorff Distance (HD) for multiple isodose levels in both the phantoms and patients. Biologically corrected dose distributions were calculated and compared across plans using two RBE strategies: (1) RBE 1.1 , corrected assuming uniformly 10% higher effectiveness of the proton plans, and (2) RBE LET , calculated using a LET/RBE research script (Eclipse treatment planning system), with the LET distributions normalized to achieve an average RBE LET = 1.1 within the The CI improved in the phantom plans with up to six beams while no further benefit was observed with larger number of beams. The 3-beam plans showed high-LET and high- RBE regions at the distal edge of each beam while the 12- beam plans resulted in a more uniform distribution of LET and RBE, with the RBE closer to 1.1 (Fig. 1). Similar patterns were seen in both phantoms and patients. Compared to the 3-beam plans, the 12-beam plans had approximately 2 mm lower HD between the target and the 45 Gy isodose surfaces in all directions. The plans with increased number of beams had a more uniform HD between the target and intermediate dose surfaces (10 and 20 Gy) across all directions while the corresponding HD for the 3-beam plans were highly depending on beam direction. The 3-beam plans had the smallest low-dose ‘bath’ (at 2 and 5 Gy; Fig. 2). target. Results

Conclusion The physical conformity at high-dose levels in both phantom and patient plans improved when adding up to six fields, suggesting a potential for proton arc therapy. Increasing the number of beams also resulted in more uniform biological dose distributions, due to the reduced high-LET volumes. However, increasing the number of beams also enlarged the low-dose ‘bath’ (<5 Gy) with possible implications for secondary cancer risks. PO-0935 Implementation of an in-house solution for motion management-based treatment planning D. Dechambre 1 , A. Delor 1 , F. Vanneste 1 , X. Geets 1 1 Cliniques Universitaires Saint-Luc, Radiotherapy department, Brussels, Belgium Purpose or Objective No commercial system provides an approach to 4DCT treatment planning with mid-ventilation (MidV) or mid- position (MidP) as respiration-induced target motion strategies. Therefore clinics have to develop their own and often complicated solution using mostly third party software (e.g Reggui). The aim of this study was to evaluate the feasibility of using an in-house treatment planning system-based (TPS) script to replace, simplify and automatize our current and complex MidP workflow. Material and Methods Scripting was edited in RayStation (RS) system v8 using IronPython v2.7. A graphical user interface (GUI) was designed using XAML which allows for flexible and interactive data visualization. Prior GTV delineation was performed on a single 4DCT phase. Hybrid deformable registration was processed with the remaining phases and the resulting displacement fields were used to map the tumour contour. Given the impossibility to reconstruct the MidP CT from RS using diffeomorphic deformations, MidV-like approach was developed where the selected 4DCT phase minimizes the 3D vector between GTVs and MidP centers of mass (CoM). Chosen MidV_GTV is then transferred onto the time- average CT and translated to the MidP CoM coordinates, correcting for the hysteresis movement within the GTV rigidity hypothesis. Corresponding 4DCT phase was used for the organs-at-risk delineation. To validate the methodology, twenty consecutive lung SBRT cases treated in our institution were retrospectively