ESTRO 2021 Abstract Book

S1493

ESTRO 2021

Conclusion 3D positional verification of small lung lesions with modest motion, using markerless real-time tracking of 2DkV projection images acquired during irradiation was feasible. However, template matching and triangulation could not report the tumor position for on average 29% of the time, indicating that improvements are needed. When clinically necessary, geometric margin reduction may be feasible. Imaging and delivery of a single 34Gy fraction using 2400MU/min FFF VMAT, with 3 CBCT scans, was fast. PO-1766 Validation of layer rescanning techniques for mediastinal treatments in proton therapy with PBS G. Castiglione Minischetti 1 , M. Schwarz 2 , E. Engwall 3 , F. Fracchiolla 2 1 University of Trento, Department of Physics, Trento, Italy; 2 Proton Therapy Center, Trento Hospital, Trento, Italy; 3 RaySearch Laboratories AB, RaySearch, Stockholm, Sweden Purpose or Objective To analyse different layered rescanning techniques for the mitigation of interplay effects in the treatment of mediastinal lesions with proton beam scanning. Materials and Methods We included in this study 4 cases treated in our centre (1 lymphoma and 3 thymomas) and a plan generated on an anthropomorphic phantom. A free breathing CT (3DCT) and a 10-phase 4DCT were available for each case. For one patient, 2 4DCTs acquired during the treatment course were available too. In addition to the treatment delivery without rescanning, we tested the layered repainting techniques with number of rescan per energy layer: • Fixed (3 to 7 rescans) • Defined as a function of the maximum MU per layer (MML) (threshold range 1-9) • Defined as a function of the maximum MU per spot (MMS) (threshold range 0.04-0.3). Interplay effects were evaluated considering the CTVs coverage (D 95 ) and hotspot (D 1 ) on 4D dose distributions (4DDD) computed on our treatment planning system (RayStation v8B-DTK). 4DDD were calculated considering: • The delivery time structure (scanning velocity: 250 cm/s; energy layer switching: 1.2 s; dose rate: 150 MU/s); • The CTV motion, described by the 10 phase 4DCT; the 4DDD are obtained summing up the doses from each phase of the 4DCT on the 3DCT via deformable registration. For each delivery technique, 50 4DDD were computed combining 5 breathing periods (2-6 s) and considering every 4DCTs’ phase as the initial phase. The voxelwise worst dose distributions and the worst case scenarios were used for plan evaluation. The phantom plans were used to perform measurements with Gafchromic films (EBT3) on a moving platform able to simulate a sinusoidal motion. The measurement with no motion was considered as the reference dose distribution. The differences between the 2D dose distributions were evaluated with the Gamma passing rate (γPR) (2%2mm and 2%1mm). Results The rescanning techniques increased the D 95 and reduced the D 1 as compared to the no rescanning case. These changes are summarized in figure 1 for the voxelwise worst dose distributions. The same changes were observed for the worst scenarios where D 95 increases ranged between 0.44% and 2.74% for fixed rescanning. For fixed rescanning, 4 rescans reached the best compromise between quality of the dose distribution and delivery time. The results for MMS were strongly dependent on the threshold value and prompted a new implementation of this feature. The results were confirmed when robustness analysis was performed on repeated 4DCT acquired during the treatment. The γPR from the phantom irradiations are summarized in table 1 . MML and fixed showed always an increase in γPR with respect to the no rescanning case. MMS showed lower γPRs.