Use of deformed intensity distributions for on-line modification of image-guided IMRT to account for interfractional anatomic changes
Use of deformed intensity distributions for on-line modification of image-guided IMRT to account for interfractional anatomic changes
复制标题
DOI：
10.1016/j.ijrobp.2004.11.033
复制
发表时间：
2005-03-15
影响因子：
7
通讯作者：
Dong, L
中科院分区：
文献类型：
--
作者：
Mohan, R;Zhang, XD;Dong, L
目的：最近的影像学研究表明，由于日常定位不确定性以及生理和临床因素，每天以及在放疗过程中，解剖结构可能会发生显著变化。有许多策略可以最大限度地减少此类更改，减少其影响或纠正它们。迄今为止的措施包括改善外部和内部解剖结构的固定或基于射野或超声图像调整位置。也许最准确的方法是使用CT图像引导的放射治疗，其可能性范围从基于日常CT图像的简单设置校正到在线近实时调强放射治疗（IMRT）重新规划。此外，还有许多中间的可能性。在本文中，我们报告了一种这样的中间方法的发展，该中间方法通过基于如在射束的眼睛视图中看到的解剖结构的变形来变形每个射束的强度分布来考虑解剖结构变化。方法和材料：强度分布变形基于从当前图像相对于参考图像（例如，预处理CT扫描）。首先，基于参考CT图像生成参考IMRT计划。使用室内CT采集每个部分的新CT图像。获得新图像的解剖结构轮廓。(For在这篇文章中，这些轮廓是手工绘制的。当实施图像引导的IMRT方法时，后续图像上的解剖结构轮廓将可能通过自动或半自动手段获得。这可以通过例如首先使原始CT图像变形以匹配今天的图像，然后使用相同的变形变换将原始轮廓映射到今天的图像来实现。然后使每个射束的参考强度分布变形，使得从参考图像提取的解剖结构（目标组织和正常组织两者）与参考强度分布之间的射束眼视图内的投影几何关系与从今天的图像导出的解剖结构与新变形的强度分布之间的对应关系相同（或尽可能接近）。为了验证使用变形的强度分布计算的剂量分布与原始强度分布相比对于治疗是否是可接受的，变形的强度被变换成叶序列，其然后被用于计算预期被递送的强度和剂量分布。还计算了相应的剂量-体积直方图、剂量-体积指数和剂量-反应指数。将这些数据与（a）从应用于原始图像的原始治疗计划、（B）从应用于今天的图像的原始治疗计划以及（c）从基于今天的图像设计的新的成熟IMRT计划导出的相应数据进行比较。根据解剖学变化的程度，使用基于原始计划CT设计的IMRT计划来治疗当前部分可能导致与预期剂量分布相比的显著差异。与基于皮肤标记或骨标志的设置相比，CT引导的设置可以在一定程度上改善剂量分布。基于当前分数的图像重新规划IMRT产生最佳的物理可交付计划（“黄金标准”）。对于作为原理证明而研究的前列腺和头颈部示例，与其他替代方案相比，基于在射束视图中看到的解剖结构在每个射束内变形强度的结果是完全重新规划的良好近似。结论：我们的初步结果鼓励我们相信，考虑到解剖结构变形的变形强度可能是一种快速产生新治疗计划的方法，基于每日CT图像的近实时线路。我们开发的方法需要应用于一组前列腺和头颈部病例的患者，以确认我们方法的有效性。(c)2005年爱思唯尔公司
Purpose: Recent imaging studies have demonstrated that there can be significant changes in anatomy from day to day and over the course of radiotherapy as a result of daily positioning uncertainties and physiologic and clinical factors. There are a number of strategies to minimize such changes, reduce their impact, or correct for them. Measures to date have included improved immobilization of external and internal anatomy or adjustment of positions based on portal or ultrasound images. Perhaps the most accurate way is to use CT image-guided radiotherapy, for which the possibilities range from simple correction of setup based on daily CT images to on-line near real-time intensity modulated radiotherapy (IMRT) replanning. In addition, there are numerous intermediate possibilities. In this paper, we report the development of one such intermediate method that takes into account anatomic changes by deforming the intensity distributions of each beam based on deformations of anatomy as seen in the beam's-eye-view.Methods and Materials: The intensity distribution deformations are computed based on anatomy deformations discerned from the changes in the current image relative to a reference image (e.g., the pretreatment CT scan). First, a reference IMRT plan is generated based on the reference CT image. A new CT image is acquired using an in-room CT for every fraction. The anatomic structure contours are obtained for the new image. (For this article, these contours were manually drawn. When image guided IMRT methods are implemented, anatomic structure contours on subsequent images will likely be obtained with automatic or semiautomatic means. This could be achieved by, for example, first deforming the original CT image to match today's image, and then using the same deformation transformation to map original contours to today's image.) The reference intensity distributions for each beam are then deformed so that the projected geometric relationship within the beam's-eye-view between the anatomy (both target and normal tissues) extracted from the reference image and the reference intensity distribution is the same as (or as close as possible to) the corresponding relationship between anatomy derived from today's image and the newly deformed intensity distributions. To verify whether the dose distributions calculated using the deformed intensity distributions are acceptable for treatment as compared to the original intensity distributions, the deformed intensities are transformed into leaf sequences, which are then used to compute intensity and dose distributions expected to be delivered. The corresponding dose-volume histograms and dose-volume and dose-response indices are also computed. These data are compared with the corresponding data derived (a) from the original treatment plan applied to the original image, (b) from the original treatment plan applied to today's image, and (c) from a new full-fledged IMRT plan designed based on today's image.Results: Depending on the degree of anatomic changes, the use of an IMRT plan designed based on the original planning CT for the treatment of the current fraction could lead to significant differences compared to the intended dose distributions. CT-guided setup compared to the setup based on skin marks or bony landmarks may improve dose distributions somewhat. Replanning IMRT based on the current fraction's image yields the best physically deliverable plan (the "gold standard"). For the prostate and head-and-neck examples studied as proof of principle, the results of deforming intensities within each beam based on the anatomy seen in the beam's eye-view are a good approximation of full-fledged replanning compared with other alternatives.Conclusions: Our preliminary results encourage us to believe that deforming intensities taking into account deformation in the anatomy may be a rapid way to produce new treatment plans on-line in near real-time based on daily CT images. The methods we have developed need to be applied to a group of patients for both prostate and head-and-neck cases to confirm the validity of our approach. (c) 2005 Elsevier Inc.