Weichert Frank, Müller Heinrich, Quast Ulrich, Kraushaar Andreas, Spilles Peter, Heintz Martin, Wilke Carsten, von Birgelen Clemens, Erbel Raimund, Wegener Dietrich
Department of Computer Science VII, Dortmund University, D 44221 Dortmund, Germany.
Med Phys. 2003 Sep;30(9):2530-6. doi: 10.1118/1.1603964.
Intravascular brachytherapy (IVB) can significantly reduce the risk of restenosis after interventional treatment of stenotic arteries, if planned and applied correctly. To facilitate computer-based IVB planning, a three-dimensional vessel model has been derived from information on coronary artery segments acquired by intravascular ultrasound (IVUS) and biplane angiography. Part I describes the approach of model construction and presents possibilities of visualization. The vessel model is represented by a voxel volume. Polygonal information about the vessel wall structure is derived by segmentation from a sequence of IVUS images automatically acquired ECG gated during pull back of the IVUS transducer. To detect horizontal, vertical, and radial contours, modified Canny-Edge and Shen-Castan filters are applied on Cartesian and polar coordinate representations of the IVUS tomograms as edge detectors. The spatial course of the vessel wall layers is traced in reconstructed longitudinal IVUS scans. By resampling the sequence of IVUS frames the voxel volume is obtained. For this purpose the frames are properly located in space and augmented with additional intermediate frames generated by interpolation. Their spatial location and orientation is derived from biplane X-ray angiography which is performed simultaneously. For resampling, two approaches are proposed: insertion of the vertices of the rectangular goal grid into the cells of a deformed hexahedral mesh derived from the IVUS sequence, and insertion of the vertices of the hexahedral mesh into the cells of the rectangular grid. Finally, the vessel model is visualized by methods of combined volume and polygon rendering. The segmentation process is verified as being in good agreement with results obtained by manual contour tracing with a commercial system. Our approach of construction of the vessel model has been implemented into an interactive software system, 3D IVUS-View, serving as the basis of a future system for intracoronary brachytherapy treatment planning being currently under development (Part II).
如果规划和应用得当,血管内近距离放射治疗(IVB)可显著降低狭窄动脉介入治疗后再狭窄的风险。为便于基于计算机的IVB规划,已根据血管内超声(IVUS)和双平面血管造影获取的冠状动脉节段信息生成了三维血管模型。第一部分描述了模型构建方法并展示了可视化的可能性。血管模型由体素体积表示。通过在IVUS换能器回撤过程中自动采集的心电图门控IVUS图像序列进行分割,得出有关血管壁结构的多边形信息。为检测水平、垂直和径向轮廓,将改进的Canny边缘滤波器和Shen-Castan滤波器应用于IVUS断层图像的笛卡尔坐标和极坐标表示,作为边缘检测器。在重建的纵向IVUS扫描中追踪血管壁层的空间走向。通过对IVUS帧序列进行重采样获得体素体积。为此,将帧正确定位在空间中,并通过插值生成额外的中间帧进行补充。它们的空间位置和方向来自同时进行的双平面X射线血管造影。对于重采样,提出了两种方法:将矩形目标网格的顶点插入从IVUS序列导出的变形六面体网格的单元中,以及将六面体网格的顶点插入矩形网格的单元中。最后,通过体素和多边形渲染相结合的方法对血管模型进行可视化。经验证,分割过程与使用商业系统进行手动轮廓追踪得到的结果高度一致。我们构建血管模型的方法已在交互式软件系统3D IVUS-View中实现,该系统作为目前正在开发的冠状动脉内近距离放射治疗治疗规划未来系统的基础(第二部分)。