ESTRO 38 Abstract book

S1078 ESTRO 38

changes between subsequent DIBHs. It is of clinical interest to be able to identify the patients where variations in anatomy between DIBHs would result in a sub- optimal treatment. The purpose of the current study was to develop a method for evaluating the intra-fraction robustness of DIBH-RT treatment plans. Two treatment plans were evaluated as a proof of principle. Material and Methods We chose two example patients from a previous DIBH study, where three repeated DIBH scans were acquired before treatment. Patient 1 had good agreement between scans and was treated in DIBH, whereas patient 2 was treated in free breathing because of large variation in tumor position between scans. The DIBH plans (2 Gy x 33, volumetric modulated arc therapy with two partial arcs) were divided into four sub-arcs, each of duration of one DIBH (~20 s), using the Eclipse Scripting Application Programming Interface (ESAPI) version 15.5 (Varian). The three DIBH scans were assigned randomly to the four sub- arcs in a simulation of all 33 fractions, to evaluate the impact of intra-fraction variation throughout the treatment course. Subsequently, a plan with corresponding field weights was created for each DIBH scan, followed by deformable registration and dose summation in Velocity version 4.0. The resulting simulated sum dose was compared to the planned dose. The workflow is illustrated in Figure 1.

phases. The structures guided the deformable image registration (DIR). Five spherical liver tumours of radius R =1-5 cm were manually delineated as CTVs in the end- exhale phase and mapped to other phases by DIR. A dose of 63 Gy RBE was prescribed to the CTVs in 15 fractions. The magnitude of interplay effects was analysed based on an experimentally validated 4D dose calculation routine including DIR and 50 interplay scenarios per plan with varying field delivery time structures [1]. The homogeneity index (HI=(D 5 -D 95 )/D prescribed ) and the percentage over- and underdosage (V 107/95 =V 107 +100-V 95 ) of the CTV served as evaluation criteria. Results All workflow steps from data import to interplay simulations were successfully completed for the v4D CT. The simulated CT data enabled fast, threshold-based segmentation of 48 anatomical structures in all 4D CT phases. Thus, 4D studies could be performed with perfect contouring, minimal DIR uncertainties and realistic image quality without the influence of CT artefacts. Fig. 1 shows exemplarily two 4D dose distributions. While the HI increased with growing target volume from 11.5%±3.0% (average value ± standard deviation of all scenarios) for R =1 cm to 22.3%±2.4% for R =5 cm, V 107/95 did not reproduce this trend. V 107/95 was maximal for the largest CTV (39.9% ± 5.4%), but the smallest value was observed for R =3 cm (22.3% ± 7.6%).

Conclusion The results demonstrate the feasibility of 4D CT simulation based on XCAT phantoms for interplay evaluations. The manifestation of investigated interplay effects exhibited complexity, without clear correlation between the interplay level and the target size. Studies on larger scale must show whether a categorisation to influencing factors such as tumour motion or size into different treatment groups is possible and whether it can replace time consuming, individual interplay evaluations. The application of v4D CTs can facilitate these studies and furthermore allows investigation of mitigation strategies in a systematic approach. [1] Pfeiler et al. Z Med Phys 2018;28:121–33. EP-1975 Intra-fraction robustness evaluation of deep inspiration breath hold radiotherapy for lung cancer K. Håkansson 1 , M. Josipovic 1 , I.R. Vogelius 1,2 , G. Persson 2,3 , C. Behrens 3 1 Rigshospitalet- Copenhagen University Hospital, Department of Oncology- Section of Radiotherapy, Copenhagen, Denmark ; 2 Faculty of Health Science- University of Copenhagen, Department of Clinical Medicine, Copenhagen, Denmark ; 3 Herlev Hospital- Copenhagen University Hospital, Department of Oncology, Copenhagen, Denmark Purpose or Objective Deep inspiration breath hold (DIBH) for radiotherapy (RT) of lung cancer has recently been shown to be feasible and can reduce dose to the heart and lungs. Robustness may be an issue, however, and built in tools in the treatment planning systems typically do not account for anatomical

Results The ESAPI script could successfully divide a treatment plan into four sub-arcs. The robustness evaluation method showed small differences in target and organs at risk (OAR) doses between the original plan and the Velocity sum for patient 1. For patient 2, the dose differences were similar for OARs but large for the target. The difference was most pronounced for the coverage of the T-site lesion – V95% was 97.1% for PTV-T in the original plan but only 71.2% in the Velocity sum plan (Table 1).

Conclusion We demonstrated a method for dose-volume assessment of the impact of anatomical changes across consecutive DIBHs on a treatment plan. With the proposed method, a patient could be identified where the DIBH technique was sensitive to anatomical changes between repeated DIBHs