SU‐FF‐T‐05: 4D Dosimetry in Radiotherapy Using Optical Flow and Monte Carlo Techniques

Abstract

Purpose: Thoracic tumors present difficulties in dosimetry due to the respiratory motion and inhomogeneity. This study demonstrates that the combination of Monte Carlo dose calculation and 4D dose handling using deformable image registration provides a way to accurately calculate 4D dosimetry for such cases. Method and Materials: Breath‐hold and 4D CT images of lung cancer cases were used in this proof‐of‐concept study. A 4D CT image set of a phantom is also included in this study to compare the calculated 4D dose with measurements. A 3‐D conformal treatment plan was designed on the maximum expiration image volume. The treatment plan isocenter was placed at the tumor centre of the maximum expiration image set. BEAMnrc, a Monte Carlo simulation user code based on EGSnrc, a package for the Monte Carlo simulation of coupled electron‐photon transport, was used in the dose calculations for multiple respiration phases. The 3D optical flow program, which uses image‐intensity‐gradient based deformable registration algorithm, was applied to the calculation of the displacement matrices between the CT images of different respiration phases. The matrices were then used to map the dose distributions of all the respiration phases to the total dose. Results: The 4D phantom study showed that the calculated doses and the measurements agreed with in 2% in the center of the tumor insert. In a left lung tumor case, the tumor motion range was about 0.5 cm. Although the dose at the center of the tumor did not change much, the 4D integrated isodose lines moved in the direction of the tumor motion comparing to the static plans, which is consistent to the expectation. Conclusion: 4D dosimetry using Monte Carlo simulation and deformable image registration provides a promising way to handle thoracic cancer radiation dose calculation.