基于蒙特卡罗模拟的离子CT图像重建及相对阻止本领误差研究?

Ion beam radiotherapy is increasingly being used for tumor treatment. Currently, the use of traditional X-ray computed tomography (CT) for treatment planning of ion-beam radiotherapy has significant relative stopping power (RSP) errors. A more ideal approach is to directly use ion beams to generate patient images for treatment planning so as to avoid RSP conversion and reduce RSP errors. In this study, the Monte Carlo program Geent4/Gate was adopted to establish an ion-CT simulation platform, designed two sets of ion CT systems, ideal and real, and reconstructed images using the maximum likelihood method and ASD-POCS algorithm. The effects of the ideal and real settings, multiple energies, and different ion types on the RSP error of phantom reconstruction were investigated. The results show that the relative error of RSP for 330MeV protons in both the ideal and real settings was less than 1.547%, and the RSP reconstruction error in the ideal settings was much smaller than that in the real settings. The RSP reconstruction error of each material under realistic settings is close to three times that under ideal settings. The relative errors of RSP for protons decrease with the increase of incident energy. The relative errors of RSP were the biggest at 230MeV, and were2.855%2.468%1.653%and 2.553% in sulfur, phosphorus, carbon, and calcium materials. The RSP relative error reached its minimum at 330MeV, with 0.181%, 0.351%, 0.250%, and 0.245% in sulfur, phosphorus, carbon, and calcium materials. At energy of 330MeV/u for carbon ions , the RSP relative errors in sulfur, phosphorus, carbon, and calcium materials were 0.060%, 0.281%, 0.150%, and 0.082%, respectively, all within 0.281%. And the RSP relative  errors were much smaller than those of protons under the same conditions. Thus, compared with proton CT, carbon ion-beam CT seems more possible to provide accurate RSP data for treatment planning for ion-beam radiotherapy..