基于模拟退火法的医用电子加速器6MV X射线能谱重建  

Reconstruction of 6MV X-ray Spectra of Medical Linear Accelerator Based on Simulated Annealing Algorithm

作　　者：刘娟[1] 周正东[1] 陈元华[1] 王东东[1] 

机构地区：[1]南京航空航天大学核科学与工程系,南京211106

出　　处：《中国生物医学工程学报》2013年第4期385-389,共5页Chinese Journal of Biomedical Engineering

摘　　要：根据测量的中轴百分深度剂量(PDD)以及Monte-Carlo模拟的单能光子PDD,研究基于模拟退火(SA)算法重建医用电子加速器6MV X射线能谱的方法。在优化过程中,选择60个能量间隔,对应不同的相对权重,选择目标函数为重建PDD(即各相对权重与Monte-Carlo模拟单能光子PDD的乘积)与实测PDD之间的相关系数,运用模拟退火进行优化得到的最优解就是加速器的能谱。为了验证算法的有效性,对加速器治疗头进行Monte-Carlo模拟,得到从治疗头出射的6 MV光子能谱。实验结果表明,计算能谱与Monte-Carlo模拟能谱在能谱形状、峰值能量方面一致;同时,根据重建能谱获得的PDD与实测PDD保持高度一致,均方根误差为1.56×10-4。上述实验结果表明,基于模拟退火算法重建光子能谱有效可靠。A method based on simulated annealing （SA） algorithm to reconstruct 6MV X-ray energy spectra from medical linear accelerators was investigated in this work, utilizing both measured percentage depth doses （PDD） and Monte-Carlo simulated central axis PDD of mono-energy photons. As for the SA optimization, 60 energy bins were employed with different relative weights and the objective function was chosen as the correlation coefficient between reconstructed PDD （the products of relative weights and Monte-Carlo simulated PDD） and measured PDD. The optimal solution calculated by SA algorithm was exactly the spectrum of the accelerator. To verify the proposed method, a Monte-Carlo simulation of the accelerator treatment head was performed to calculate the 6MV photon energy spectrum for comparison. Results showed that the reconstructed spectrum was in good agreement with the Monte-Carlo simulated spectrum in both shape and peak position; the derived PDD from reconstructed spectrum were highly consistent with the measured PDD with a root-mean- square error of 1.56 x 10 -4. Above experimental results indicated that reconstructing photon spectra based on SA algorithm with measured PDD was effective and reliable.

关 键 词：模拟退火 MONTE-CARLO模拟 X射线能谱 百分深度剂量 

分 类 号：R318[医药卫生—生物医学工程] TP391[医药卫生—基础医学]