Tecnologías de la Información y las Comunicaciones Department, Universidad Politécnica de Cartagena, Campus Muralla del Mar Antiguo Cuartel de Antigones, s/n 30202, Cartagena, Spain.
J Digit Imaging. 2011 Dec;24(6):999-1009. doi: 10.1007/s10278-011-9376-z.
Image processing turns out to be essential in the planning and verification of radiotherapy treatments. Before applying a radiotherapy treatment, a dosimetry planning must be performed. Usually, the planning is done by means of an X-ray volumetric analysis using computerized tomography, where the area to be radiated is marked out. During the treatment phase, it is necessary to place the patient under the particle accelerator exactly as considered in the dosimetry stage. Coarse alignment is achieved using fiduciary markers placed over the patient's skin as external references. Later, fine alignment is provided by comparing a digitally reconstructed radiography (DRR) from the planning stage and a portal image captured by the accelerator in the treatment stage. The preprocessing of DRR and portal images, as well as the minimization of the non-shared information between both kinds of images, is mandatory for the correct operation of the image registration algorithm. With this purpose, mathematical morphology and image processing techniques have been used. The present work describes a fully automatic method to calculate more accurately the necessary displacement of the couch to place the patient exactly at the planned position. The proposed method to achieve the correct positioning of the patient is based on advanced image registration techniques. Preliminary results show a perfect match with the displacement estimated by the physician.
图像处理在放射治疗的计划和验证中至关重要。在实施放射治疗之前,必须进行剂量学计划。通常,该计划是通过计算机断层扫描的 X 射线体积分析来完成的,其中标记出要辐射的区域。在治疗阶段,需要将患者精确地放置在粒子加速器下,就像在剂量学阶段所考虑的那样。通过在患者皮肤上放置作为外部参考的信托标记来实现粗略对准。之后,通过比较计划阶段的数字重建射线照相术(DRR)和加速器在治疗阶段捕获的射野影像来提供精细对准。为了正确运行图像配准算法,必须对 DRR 和射野影像进行预处理,并最小化两种影像之间的非共享信息。为此,已经使用了数学形态学和图像处理技术。本工作描述了一种全自动方法,可以更准确地计算将患者精确放置在计划位置所需的床位移。为实现患者的正确定位,提出了一种基于先进的图像配准技术的方法。初步结果显示与医生估计的位移完全匹配。