Martonen T B, Hwang D, Guan X, Fleming J S
Experimental Toxicology Division, U.S. EPA, Research Triangle Park, North Carolina 27711, USA.
J Nucl Med. 1998 Apr;39(4):745-50.
A supercomputer code that describes the three-dimensional branching structure of the human lung has been developed. The algorithm was written for the Cray C94. In our simulations, the human lung was divided into a matrix containing discrete volumes (voxels) so as to be compatible with analyses of SPECT images. The matrix has 3840 voxels. The matrix can be segmented into transverse, sagittal and coronal layers analogous to human subject examinations. The compositions of individual voxels were identified by the type and respective number of airways present. The code provides a mapping of the spatial positions of the almost 17 million airways in human lungs and unambiguously assigns each airway to a voxel. Thus, the clinician and research scientist in the medical arena have a powerful new tool to be used in imaging analyses. The code was designed to be integrated into diverse applications, including the interpretation of SPECT images, the design of inhalation exposure experiments and the targeted delivery of inhaled pharmacologic drugs.
一种描述人类肺部三维分支结构的超级计算机代码已被开发出来。该算法是为克雷C94编写的。在我们的模拟中,人类肺部被划分为一个包含离散体积(体素)的矩阵,以便与单光子发射计算机断层扫描(SPECT)图像分析兼容。该矩阵有3840个体素。该矩阵可以分割成类似于人体检查的横向、矢状和冠状层。通过存在的气道类型和各自数量来识别各个体素的组成。该代码提供了人类肺部近1700万个气道空间位置的映射,并明确地将每个气道分配到一个体素。因此,医学领域的临床医生和科研人员有了一个可用于成像分析的强大新工具。该代码被设计用于集成到各种应用中,包括SPECT图像的解读、吸入暴露实验的设计以及吸入药物的靶向递送。