中文摘要：

目的研究基于CT-MRI融合图像的三维重建技术在颅底外科中的应用。方法采集25例鼻腔鼻窦占位病变患者的鼻窦增强CT及头颅MRI薄层扫描图像数据，使用自行编制的三维重建软件FitMeDICOMviewer0．15．1对所采集数据进行处理，包括用眼球及枕骨大孔中心点进行CT-MRI的点配准，利用该软件进行融合、分割和三维可视化重建，采用虚拟内镜下观察法、体绘制分割法、自动空间测量法及人工空间测距法对重建后的模型进行操作、观察及测量。结果融合后图像中的重要解剖标志吻合度好。分割的图形边界清晰、完整。重建的结构显示完整、连续、表面光滑。模拟内镜经鼻的观察方式可从鼻腔内部显示颅底各组织的立体结构，三维旋转及透明度调节流畅。体绘制分割后的颅骨残端边界清楚、边缘圆滑，可同时显示骨性标志与颅内软组织模型。测量25例患者肿物模型最大径线自动测量法为（32．007±15．311）mm，人工测量法为（30．240±15．169）mm，差异有统计学意义（t=8．409，P〈0．05）。结论利用眼球及枕骨大孔中心点进行点配准的方法可用于头颅CT—MRI图像配准。基于CT—MRI融合图像三维重建出的模型可以准确地反映软组织的大小，通过自动测量法可以更好地进行模型三维测量。重建模型可通过虚拟内镜或体绘制分割来实现由鼻腔内部观察或由外向内观察颅底结构的空间位置关系。将CT—MRI融合图像应用于颅底外科三维重建中，可进行手术入路规划、手术风险评估及空间测茸．并为导航图像的市体显示提供参考．

英文摘要：

Objective To study the application of three-dimensional reconstruction technique based on CT-MRI fusion in skull base surgery. Methods To acquire the thin layer CT scan and MRI scanned images, to achieve image registration, fusion, segmentation and 3D visualization by using self-preparation software, to operate, observe and measure models by using methods of endoscopic observation, volume rendering segmentation, automatically and manually measure. Results The center of the eye and foramen magnum in CT-MRI were used as point registration. Good coincidence of important anatomic landmarks were formed in the image fusion. The boundary of spirted graphical was clear and complete. The models showed a complete, continuous, smooth surface. Virtual endoscopy could display the inside three-dimensional structures of skull from nasal with fluent operations of rotation and transparency. The boundary of skull stump segmented after volume rendering segmentation was clear and smooth, and it coula show bone signs and soft tissue models together. Cooperation of automatic measurement method [ （32. 007 -＋ 15.311 ） mm ] and the manual measurement method [ （30. 240 ＋ 15. 169）mm ] for measuring the maximum diameters of the tumor model, the difference was significant （ t = 8. 409, P 〈 0.05 ）. Conclusions The method of selecting the center of the eye and foramen magnum in point matching is scientific, simple and easy to operate. The models reconstructed based on CT-MRI fusion images can accurately reflect the size of the soft tissue and be better measured through the automatic measurement. Reconstruction models can be observed through the way of virtual endoscopic within the nasal cavity or volume rendering segmentation from outside to inside to frustrate the relationship of skull structures. Three-dimensional reconstruction techniques based on CT-MRI fusion in skull base surgery can be used to plan surgical approach, to assess the risk of surgery and to achieve space measurements, and it laid the foundation for the thr