Temporo-spatial IMRT optimization: concepts, implementation and initial results
Temporo-spatial IMRT optimization: concepts, implementation and initial results
复制标题
DOI：
10.1088/0031-9155/50/12/004
复制
发表时间：
2005-06-21
影响因子：
3.5
通讯作者：
Bortfeld, T
中科院分区：
文献类型：
--
作者：
Trofimov, A;Rietzel, E;Bortfeld, T
随着最近4D-CT的可用性，呼吸期间关于内部器官运动的信息的准确性已经显著提高。我们调查的效用器官的运动信息在调强放射治疗计划，使用内部原型优化系统。比较了四种方法：（1）基于运动信息，利用优化的余量进行规划;（2）运动核方法，其中通过对针对解剖结构的不同实例计算的影响矩阵进行时间加权平均来实现对从笔形射束到移动目标的剂量存款的更准确描述（4D-CT数据的子集，对应于运动的各个阶段）或通过笔形波束核与描述目标运动的概率密度函数的卷积;（3）最佳选通，或利用针对解剖结构的每个实例独立优化的射束强度图进行跟踪;以及（4）利用针对解剖结构的所有实例同时优化的射束强度图进行最佳跟踪。优化是基于梯度技术，可以处理物理（剂量体积）和等效的均匀剂量约束。优化需要从相位到相位的体素映射，以便在个体体素移动时对其剂量进行评分。结果表明，与其他方法相比，边缘扩展有一个显着的缺点，大大增加了病人的积分剂量。虽然门控或跟踪导致最佳剂量构象的目标，前者延长治疗时间，后者显着复杂的输送程序。与最佳跟踪或门控相比，运动核方法不提供剂量测定优势，但可能导致更有效的递送。门控与“运动核”或裕度扩展方法的组合将增加占空比，并且可以在递送过程的复杂性和对靶的剂量适形性方面提供最有效的解决方案。
With the recent availability of 4D-CT, the accuracy of information on internal organ motion during respiration has improved significantly. We investigate the utility of organ motion information in IMRT treatment planning, using an in-house prototype optimization system. Four approaches are compared: (1) planning with optimized margins, based on motion information; (2) the,motion kernel' approach, in which a more accurate description of the dose deposit from a pencil beam to a moving target is achieved either through time-weighted averaging of influence matrices, calculated for different instances of anatomy (subsets of 4D-CT data, corresponding to various phases of motion) or through convolution of the pencil beam kernel with the probability density function describing the target motion; (3) optimal gating, or tracking with beam intensity maps optimized independently for each instance of anatomy; and (4) optimal tracking with beam intensity maps optimized simultaneously for all instances of anatomy. The optimization is based on a gradient technique and can handle both physical (dose-volume) and equivalent uniform dose constraints. Optimization requires voxel mapping from phase to phase in order to score the dose in individual voxels as they move. The results show that, compared to the other approaches, margin expansion has a significant disadvantage by substantially increasing the integral dose to patient. While gating or tracking result in the best dose conformation to the target, the former elongates treatment time, and the latter significantly complicates the delivery procedure. The,motion kernel' approach does not provide a dosimetric advantage, compared to optimal tracking or gating, but might lead to more efficient delivery. A combination of gating with the 'motion kernel' or margin expansion approach will increase the duty cycle and may provide one with the most efficient solution, in terms of complexity of the delivery procedure and dose conformality to the target.