NONHOMOGENEOUS ANALYSIS OF 3-DIMENSIONAL TRANSMURAL FINITE DEFORMATION IN CANINE VENTRICULAR MYOCARDIUM
NONHOMOGENEOUS ANALYSIS OF 3-DIMENSIONAL TRANSMURAL FINITE DEFORMATION IN CANINE VENTRICULAR MYOCARDIUM
复制标题
DOI：
10.1016/0021-9290(91)90287-w
复制
发表时间：
1991-01-01
影响因子：
2.4
通讯作者：
OMENS, JH
中科院分区：
文献类型：
--
作者：
MCCULLOCH, AD;OMENS, JH
开发了一种新方法来分析犬心室肌有限变形的跨壁分布，而无需假设壁的有限体积中的应变是均匀的。双线性三次或双线性二次有限元的三维节点几何参数通过最小二乘拟合到植入左心室自由壁的12-18个不透射线标记的测量坐标。对于六只狗的心脏，拟合的平面内坐标的均方根误差在舒张末期参考状态下为0.079-0.556 mm，在收缩末期为0.142-0.622 mm。在径向坐标上，相应的误差范围为：舒张末期0.042- 0.264mm，收缩末期0.106- 0.279mm。平滑连续的壁应变的跨壁配置文件计算的元素变形在心动周期从end-martocole到end-systole显示出良好的协议与传统的均匀分析的离散结果。使用厚壁不可压缩圆柱体的运动学，由于心肌变形的非均匀性，整体绝对误差被发现在新的分析中减少了30-35%的典型实验参数。总体相对误差也有所降低（从23%降至20%）。由于重建的标记坐标中的测量误差通过拟合过程在空间上被平滑，所以计算的变形中的噪声也被显著衰减，并且可以以改进的精度获得三维应变分量的跨壁梯度。因此，受跨壁应力和应变分布影响的生理因素，如心肌血流量，缺血和肥大，可以更好地理解。
A new method has been developed for analyzing transmural distributions of finite deformation in canine ventricular myocardium without the need to assume that the strain in a finite volume of the wall is homogeneous. The three-dimensional nodal geometric parameters of bilinear-cubic or bilinear-quadratic finite elements are fitted by least squares to the measured coordinates of 12-18 radiopaque markers implanted in the left ventricular free wall. For six dog hearts, root-mean-squared errors in the fitted in-plane coordinates ranged from 0.079-0.556 mm in the end-diastolic reference state and 0.142-0.622 mm at end-systole. The corresponding error ranges in the radial coordinate were 0.042-0.264 mm at end-diastole and 0.106-0.279 mm at end-systole. Smoothly continuous transmural profiles of wall strain computed as the element deformed during the cardiac cycle from end-diastole to end-systole showed good agreement with the discrete results of conventional homogeneous analysis. Using the kinematics of a thick-walled incompressible cylinder, overall absolute errors due to the non-homogeneity of myocardial deformation were found to be reduced in the new analysis by 30-35% for typical experimental parameters. Overall relative errors were also reduced (from 23 to 20%). Since measurement errors in the reconstructed marker coordinates were spatially smoothed by the fitting procedure, noise in the computed deformations was also substantially attenuated, and transmural gradients of three-dimensional strain components could be obtained with improved accuracy. Hence physiological factors affected by transmural stress and strain distributions, such as myocardial blood flow, ischemia and hypertrophy, may be better understood.