摘要目的:研究瓦里安高分辨率(HD120)多叶准直器叶片位置误差对肺部肿瘤体部立体定向放疗(SBRT)的剂量学影响。方法:选取江苏省肿瘤医院9例已完成治疗的基于HD120调制的肺部肿瘤SBRT计划作为参考计划，采用基于Varian Eclipse 15.6治疗计划系统自带应用程序接口的自编程序修改计划参数，分别生成包含叶片位置同向系统误差、异向系统误差和随机误差的误差模拟计划。计算各个模拟计划与参考计划剂量学指标的偏差并做回归分析，以此评估HD120的三类叶片位置误差对SBRT的剂量学影响。结果:计划靶体积D 99%、D 2cm和双肺V 5Gy均随同向系统误差绝对值增大呈抛物线型降低，二次回归分析后得到一阶敏感度为(-0.06~-0.26)%/mm，二阶敏感度为(-0.55~-1.17)%/mm 2( R2＝0.96~0.99， P<0.01)，其中计划靶体积D 99%最大变化为-3.13%。对异向系统误差和随机误差的影响做线性回归分析，得出适形指数偏差随这两种误差变化的敏感度分别为25.16%/mm ( R2＝0.98， P<0.01)和-4.84%/mm ( R2＝0.99， P<0.01)，其余剂量学偏差随异向系统误差变化的敏感度范围为(4.80~5.12)%/mm ( R2＝0.96~0.98， P<0.01)，随随机误差变化的敏感度范围为(-0.47~-1.01)%/mm ( R2＝0.96~0.99， P＝0~0.02)。 结论:基于HD120调制的肺部肿瘤SBRT计划的剂量学偏差对叶片位置异向系统误差极为敏感，但对随机误差敏感性较弱。叶片位置同向系统误差会导致靶区覆盖剂量一定程度降低，因此临床工作中需严格控制SBRT计划执行时HD120叶片位置的系统误差。

abstractsObjective:To evaluate the effect of leaf position error of Varian high-definition multi-leaf collimator (HD120) on the dosimetry of stereotactic body radiation therapy (SBRT) for lung tumors.Methods:Nine SBRT plans based on HD120 for lung tumors were selected as the reference plans. The parameters of the plans were modified by the in-house program based on the Varian Eclipse 15.6 scripting application program interface to generate the simulation plans with three types of leaf position errors including the isotropic systematic error, the anisotropic systematic error and the random error, respectively. Then, the dosimetric metric deviation between each simulation plan and the corresponding reference plan was calculated and regression analysis was performed to evaluate the dosimetric effect of three types of leaf position errors of HD120 on SBRT.Results:The planning target volume (PTV) D 99%, D 2cm and V 5Gy of double lungs were decreased quadraticly with the increase of the absolute value of the isotropic systematic error. The first-order sensitivity was -0.06%/mm to -0.26%/mm, and the second-order sensitivity was -0.55%/mm 2 to -1.17%/mm 2 ( R2=0.96-0.99, P<0.01). The maximum change of PTV D 99% was -3.13%. The linear regression analysis of the effects of the anisotropic systematic error and random error showed that the sensitivity of CI was 25.16%/mm ( R2=0.98, P<0.01) and -4.84%/mm( R2=0.99, P<0.01), and the sensitivity of other dosimetric deviations with the anisotropic systematic error was 4.80%/mm to 5.12%/mm ( R2=0.96-0.98, P<0.01), whereas the sensitivity with the random error was -0.47%/mm to -1.01%/mm ( R2=0.96-0.99, P=0-0.02). Conclusions:The dosimetric deviation of SBRT plan based on HD120 for lung cancer is highly sensitive to the anisotropic systematic error of leaf position, but less sensitive to the random error. In addition, the isotropic systematic error of leaf position will lead to the decrease of target coverage to a certain extent. Consequently, it is necessary to strictly control the systematic error of HD120 leaf position in the implementation of SBRT plan in clinical work.