ESTRO 2021 Abstract Book

S1259

ESTRO 2021

Figure 2: Dose differences in terms of DVH parameters for gross target volume (GTV) and organs at risk (OARs).

Conclusion The dosimetric impact of daily anatomical variations were shown to be dominant over those of respiration in SBRT for peripheral lung cancer.

PO-1535 Deformable image registration uncertainty for dose accumulation of proton therapy for H&N tumors F. Amstutz 1,2 , P.G. D'Almeida 1,3 , F. Albertini 1 , L. Nenoff 1,2 , D.C. Weber 1,4,5 , A.J. Lomax 1,2 , Y. Zhang 1 1 Paul Scherrer Institute, Center for Proton Therapy, Villigen, Switzerland; 2 ETH Zurich, Departement of Physics, Zurich, Switzerland; 3 ETH Zurich, Departement of Information Technology & Electrical Engineering, Zurich, Switzerland; 4 University Hospital Zurich, Departement of Radiation Oncology, Zurich, Switzerland; 5 University Hospital Bern, Departement of Radiation Oncology, Bern, Switzerland Purpose or Objective Although proton therapy (PT) can achieve highly conformal treatment plans with good normal tissue sparing, rigid and non-rigid patient changes can impact the actually delivered dose distribution. As such, regular adaption of the delivered dose to compensate for these changes, eventually on a daily basis, is often necessary. Deformable anatomical changes however require complex anatomic correlations between imaging data sets to accurately accumulate dose from each fraction. The aim of this work is to evaluate accumulation uncertainties due to differences between deformable image registration (DIR) methods for H&N cancer patient treatments. Materials and Methods A 6-field, intensity modulated PT plan was generated on a planning (reference) CT. This, together with 8 repeated CTs acquired on different fractions during the treatment course, were aligned rigidly in Velocity. Thereafter, three DIRs (Plastimatch Bspline & Demon, and Velocity's Bspline-based method) were applied to each and the reference CT. Planning contours were propagated with each DIR for QA of the geometric accuracy (Fig.1) DICE scores were calculated for 33 structures between the propagated and corresponding clinician define structures on each repeat CT. To evaluate dosimetric uncertainty, the planned dose was then recalculated on each repeat CT and warped back to the reference using each DIR. Accumulated dose was obtained by averaging over the warped doses of 8 fractions for each DIR, and uncertainty quantified as the max-min dose of each voxel.