ESTRO 2021 Abstract Book

S779

ESTRO 2021

Fig. 2. Spatial distribution of HDNs within leg sarcoma.

Conclusion At this moment we demonstrated the technical feasibility and safety of SFGRT. This is the first time SFGRT technique was applied in our country. The whole study includes also clinical evaluations such as tumor response by histology analysis, postoperative morbidity and disease free survival. PD-0936 A two-beams approach for organs-at-risk sparing in ocular passive proton therapy E. Fleury 1,2 , P. Trnková 3,1 , C. van Rij 1 , E. Kiliç 4 , A. Zolnay 1 , K. Spruijt 2 , J. Pignol 5 , M. Hoogeman 1,2 1 Erasmus Medical Center, Radiation Oncology, Rotterdam, The Netherlands; 2 Holland Proton Therapy Center, Radiation Oncology, Delft, The Netherlands; 3 Medical University of Vienna, Radiation Oncology, Vienna, Austria; 4 Erasmus Medical Center, Ophthalmology, Rotterdam, The Netherlands; 5 Dalhousie University, Radiation Oncology, Halifax, Canada Purpose or Objective Proton therapy for Uveal Melanoma (UM) is currently planned on a generic and simplistic ocular model without accurate 3D delineations of UM and organs-at-risk (OARs) from either CT or MRI. For treatment, only one forward planned single anterior beam is selected for dose delivery. Both are shortcomings that have potential to be optimized. The aim of this project was to explore the potential of 3D imaging-based planning using two optimal beams instead of a single one. Materials and Methods Planning CT (pCT) and clinical contours of 15 UM patients initially treated with a fractionated stereotactic radiation therapy treatment were collected. The gross tumor volume (GTV), optic nerve, anterior segment (including cornea, lens and ciliary body) and lacrimal gland were delineated on the pCT. Dose calculations were performed using our in-house proton eye dose algorithm for passive scattering, with a prescribed dose of 60 Gy(RBE). An isotropic 2.5 mm margin was set by adjusting the range and modulation width of the spread- out Bragg peak proximal and distal edges for GTV coverage. Lateral conformation was achieved by a collimator adjusted to the (GTV + 2.5 mm margin) contour. Two optimal beams for patient gazing were selected: the first one always prioritized the optic nerve (constraint D 2% =30Gy(RBE)), whereas the second one was configured for anterior segment and lacrimal gland sparing. For evaluation, patients were clustered in 2 groups (A and B) based on the minimum distance between GTV edge and optic nerve with a threshold given at 3 mm corresponding to the physics limitations in terms of distal and lateral penumbrae of the proton beams. For the 3 OARs, D 2% and D mean were evaluated for each single beam vs. the 2-beams approach, as well as D 25% for the anterior segment. Results GTV coverage was reached (D 95% > 95%) for all cases for the 3 beam configurations. An example of the proton dose distributions is shown in Figure 1. The dose distributions of the planned gazing angles were registered to the gazing angle hold by the patient during CT. Figure 1.A and 1.B display plans with a single beam, while Figure 1.C displays the dose distribution for the equally-weighted combined 2-beams. The dosimetric parameter values of both clusters are shown by boxplots in Figure 2. For cluster A, the use of 2 beams compared to the first optimal beam lowered the D 25% of the anterior segment (median: 45 and 27 Gy(RBE), respectively), while maintaining the D 2% of the optic nerve (median 1.2 Gy(RBE)). For cluster B, the use of 2 beams compared to the first optimal beam lowered the D 25% of the anterior segment (median: 38 to 21 Gy(RBE), respectively), and the D 2% of the lacrimal gland (median: 32 and 19 Gy(RBE), respectively), while maintaining the D 2% of the optic nerve (median 30 Gy(RBE)).