带有预处理的计划CT和在线CT的形变配准方法-《中国医学物理学杂志》

文章信息/Info

Title:: A preprocessing-based method for deformation registration of planned CT and online CT

Author(s):: LIU Huan; YU Gang; LU Yukun; LI Dengwang; School of Physics and Electronics, Shandong Normal University; Shandong Province Key Laboratory of Medical Physics and Image Processing Technology, Ji’nan 250000, China

Keywords:: planned CT; online CT; image registration; Demons algorithm; preprocessing; segmentation

摘要:: CT扫描中，水溶性碘造影的存在使得计划CT和在线CT图像中血管内的HU值出现非常大的偏差，从而导致计划CT和在线CT图像错配。针对该问题，本研究提出了一种基于预处理的计划CT和在线CT形变配准方法。首先，根据CT图像组织和结构的信息，利用阈值分割方法分割出血管，并将所有分割中最大的联通区域作为初始分割的强化血管；其次，利用分割得到的强化血管区域外扩5 mm，作为外扩的强化血管，并将血管用固定的HU值进行填充；最后，对完成填充后的图像利用Demons算法进行形变配准。实验结果显示本文提出的带有预处理的形变配准方法，可以较好地解决水溶性碘造影剂引起的CT错配问题。

Abstract:: During CT scanning, the existence of water-soluble iodine contrast agent makes the HU values of the blood vessels on planned CT images and online CT images deviate greatly, which leads to the mismatch between planned CT images and online CT images. Herein we propose a preprocessing-based method for the deformation registration between planned CT and online CT in order to solve the problem caused by water-soluble iodine contrast agent. Firstly, according to the information of tissues and structures on CT images, the blood vessels are segmented using threshold segmentation method, and the largest communicating region in all segmentations is taken as an initial segmented reinforcing blood vessel. Secondly, the segmented reinforcing blood vessel region is externally expanded by 5 mm as an expanded reinforcing blood vessel, and the blood vessel is filled with a fixed HU value. Finally, deformation registration of the filled image is carried out with the use of Demons algorithm. The results reveal that the deformation registration method with preprocessing can solve the problem of CT mismatch due to water-soluble iodine contrast agent.