多模态医学影像数据融合三维构建肢体肿瘤及其周围解剖整体结构

Three-dimensional reconstruction of whole tumor and peritumoral anatomical structures by multimodality medical image data fusion

目的利用多模态医学影像数据三维重建和配准融合的方法构建肢体肿瘤及瘤周解剖整体结构,明确肿瘤的局部特征和三维空间关系,为肿瘤外科治疗术前评估和计划提供客观的空间可视化信息。方法采集4例肢体肿瘤患者盆部和大腿上段CT、MRI和MRA影像数据,分别导入三维重建软件,基于CT和MRI对人体不同组织结构成像敏感度的差异,从各模态影像数据中分别提取肿瘤及其周围目标结构进行三维重建。依据体内解剖参照点对各个组织的三维结构进行配准融合,构建肢体肿瘤及瘤周解剖整体结构。结果通过CT、MRI和MRA影像数据构建出包括肿瘤及周围骨骼、肌肉、神经、血管、盆腔脏器等组织的三维整体结构,显示了肿瘤的大小、形状、部位等局部特征以及肿瘤与其周围结构的三维空间关系。结论以多模态影像数据通过三维重建和配准融合的方法可以构建肿瘤区域多种组织的三维整体结构,清楚而直观地显示肿瘤的局部特征以及与周围结构的空间解剖关系,具有个体化肿瘤术前评估和计划的应用价值。

Objective To construct whole tumor and peritumoral anatomical structures through the method of threedimensional（3D） reconstruction,registration and fusion based on muhimodality medical image data and to assess the tumor features and its spatial relationship for preoperative assessment and surgical planning for tumor treatment. Methods The image data of CT, MRI and MRA from four patients with proved pelvic and proximal thigh tumors were respectively imported into 3D reconstruction software. Based on the difference of imaging sensitivity between CT and MRI for human tissues,the tumor and peritumoral interested structures were extracted from each modality image data and further 3D reconstructed. The 3D models of all the reconstructed structures were registered and fused together according to in vivo anatomical landmarks chosen as reference points,then the whole tumor and peritumoral anatomical structures were 3I） developed. Results The whole 3D structures including tumor, muscle,nerve,vessel and pelvic or- gan were successfully reconstructed based on CT,MRI and MRA image data. The reconstructed models clearly depicted the tumor features, such as size, shape and location, and provided a vivid view of the spatial relationship among its involved regions. Conclusions The method of 3D reconstruction,registration and fusion of tumor and peritnmoral anatomical structures based on muhimodality medical image data is feasible and practical,and can directly demonstrate the tumor features in 3 D texture as well as its spatial relationship with peritumoral structures, and further offer the application value of subject-specific preoperative assessment and surgical planning for tumor treatment.