标点法MRI-CT图像融合的融合精度分析

Analysis of MRI-CT image registration accuracy by Tris-Axes Landmark

目的:研究标点法进行MRI和CT图像融合后的融合精度,以期为颅内肿瘤术后进行放射治疗的患者寻找更有效的图像融合方法及精度评价标准。方法:24例颅内肿瘤术后患者,在相同定位条件下进行MRI T1WI增强序列和术后定位CT扫描,采用Tris-Axes Landmark方法将其轴位MRI的T1WI增强序列与定位CT图像进行融合,寻找融合分数与融合误差之间的线性关系,并观察每例患者在三维空间的融合误差。结果:融合分数和融合误差之间存在较好的线性关系,相关系数r=-0.985,回归方程为y=-0.551 3x+5.351 4,据此回归方程推测,融合分数>6.61者均可满足临床融合需要。结论:标点法是放射治疗中一种较好的融合方法,能快速准确地定量评价MRI-CT融合精度,可应用于临床放射治疗中。

Objective ： To study the accuracy of MRI-CT image registration in order to find a more effective method of image fusion and a standard for the evaluation of post-operative intra-eranial tumor radio- therapy. Methods： Twenty-four patients with intra-eranial tumor underwent MR Tl-weighed sequence and post-operative location CT scan under the same location condition. The trans-axial MR Tl-weighed sequence and CT image were registered and the correlation between registration scores and errors was ana- lyzed by the method of Tris-Axes Landmark. The linear registration errors were observed in the three-di- mensional space of every patient. Results： An improved linear correlation was found between registra- tion scores and errors （ r = - 0. 985, y = - 0.551 3x ＋ 5.351 4 ）. The registration met the clinical standard when the scores exceeded 6.61. Conclusion ： Tris-Axes Landmark is a preferable registration method to be used in clinical radiotherapy for rapid and precise evaluation of MRI-CT image registration accuracy.