TH‐A‐19A‐10: Fast Four Dimensional Monte Carlo Dose Computations for Proton Therapy of Lung Cancer

Abstract

Purpose:To develop and validate a fast and accurate four dimensional (4D) Monte Carlo (MC) dose computation system for proton therapy of lung cancer and other thoracic and abdominal malignancies in which the delivered dose distributions can be affected by respiratory motion of the patient.Methods:A 4D computer tomography (CT) scan for a lung cancer patient treated with protons in our clinic was used to create a time dependent patient model using our in‐house, MCNPX‐based Monte Carlo system (“MC2”). The beam line configurations for two passively scattered proton beams used in the actual treatment were extracted from the clinical treatment plan and a set of input files was created automatically using MC2. A full MC simulation of the beam line was computed using MCNPX and a set of phase space files for each beam was collected at the distal surface of the range compensator. The particles from these phase space files were transported through the 10 voxelized patient models corresponding to the 10 phases of the breathing cycle in the 4DCT, using MCNPX and an accelerated (fast) MC code called “FDC”, developed by us and which is based on the track repeating algorithm. The accuracy of the fast algorithm was assessed by comparing the two time dependent dose distributions.Results:The error of less than 1% in 100% of the voxels in all phases of the breathing cycle was achieved using this method with a speedup of more than 1000 times.Conclusion:The proposed method, which uses full MC to simulate the beam line and the accelerated MC code FDC for the time consuming particle transport inside the complex, time dependent, geometry of the patient shows excellent accuracy together with an extraordinary speed.