ESTRO 2021 Abstract Book

S317

ESTRO 2021

OC-0420 A novel stochastic optimization method for handling uncertainty in combined proton-photon treatments S. Fabiano 1 , N. Torelli 1 , D. Papp 2 , J. Unkelbach 1 1 University Hospital Zürich, Department of Radiation Oncology, Zürich, Switzerland; 2 North Carolina State University, Department of Mathematics, Raleigh, USA Purpose or Objective Combined proton-photon treatments, where the majority of fractions are delivered with photons and only a few are delivered with protons, may represent a practical approach to optimally make use of limited proton resources. When OARs are located within or near the tumor, the optimal multi-modality treatment is a nontrivial combination of proton and photon plans. Protons hypofractionate parts of the target volume while photons are used to achieve near-uniform fractionation in dose-limiting healthy tissues. Such plans may be sensitive to range and setup errors, especially misalignments between proton and photon doses. We developed a novel stochastic optimization method to directly incorporate these uncertainties into BED-based simultaneous optimization of proton and photon plans. Materials and Methods The method considers the expected value E[b] and standard deviation sd[b] of the cumulative BED b in every voxel of a structure. For the target, a piecewise quadratic penalty function of the form (B min -(E[b]-2sd[b])) 2 + is minimized, aiming for plans in which the expected BED minus two times the standard deviation exceeds the prescribed BED B min . Analogously, ((E[b]+2sd[b])-B max ) 2 + is considered for OARs. The method is computationally efficient since the variance of the cumulative BED is the sum of proton and photon components without the need to consider all combinations of setup errors for protons and photons. Combined treatments were investigated for a spinal metastasis case. A BED 10 corresponding to 35.2 Gy in 5 fractions was prescribed to the CTV while limiting the maximum dose in the cauda to the BED 2 -equivalent of 20 Gy. Setup errors were modeled by shifting the treatment isocenter position by ±3 mm in the 3 cardinal directions. Range errors were modeled by uniformly scaling the relative stopping power by ±5%. The optimized combined treatment was compared to a naïve combination in which each modality delivers the prescribed dose per fraction to the target volume. Results Figure 1 illustrates a combined treatment that uses 1 IMPT and 4 IMRT fractions in the nominal scenario. Protons and photons delivered similar doses around the cauda to protect this serial structure through fractionation. Meanwhile, protons delivered an overproportionate dose to the remaining tumor. The optimized combination maintained tumor coverage and a good sparing of the cauda even with misalignments of 6 mm between the proton and photon fractions (Figure 2a) while achieving 42% of the healthy tissue integral dose reduction that a 5-fraction IMPT plan yields, compared to 20% for the naïve combination (Figure 2b), in the nominal scenario. Conclusion The joint planning of protons and photons may allow to optimally use limited proton resources for individual patients. Concerns about range and setup errors for safe clinical implementation of optimized proton-photon therapy can be addressed through an appropriate stochastic planning method.