Van Rietbergen B, Odgaard A, Kabel J, Huiskes R
Department of Orthopaedics, University of Nijmegen, The Netherlands.
J Orthop Res. 1998 Jan;16(1):23-8. doi: 10.1002/jor.1100160105.
It would be advantageous if the mechanical properties of trabecular bone could be directly inferred from stereomorphometric parameters. For that purpose, apparent density and mean intercept length, as measures of bone mass and directionality (fabric), are commonly correlated with the elastic characteristics of bone samples, as determined in compression tests. However, complete and accurate relationships have not yet been established in this way. This may be due not only to the occurrence of artifacts in both the stereomorphometric and the mechanical assessments but also to an inherent inadequacy of mean intercept length in characterizing the full mechanical significance of bone architecture or nonhomogeneities in trabecular tissue properties not accounted for in stereomorphometry. In this study, we introduce a computer modeling approach allowing these biases to be eliminated. With use of high-resolution three-dimensional computer reconstructions of trabecular bone specimens for stereomorphometry and for microstructural finite element models to simulate mechanical tests, unbiased comparisons become feasible. The purpose was to investigate if accurate and complete relationships can be established in this way. Four different fabric measures were considered: mean intercept length and three volume-based ones. Compliance matrices were calculated from fabric tensors, with use of the mathematical relationship proposed by Cowin for 29 vertebral whale-bone specimens. These were correlated with the compliance constants determined directly from the microstructural finite element model simulation. The nine orthotropic elastic constants of all 29 specimens were well predicted from their stereomorphometric fabric and volume fraction values, with correlation coefficients ranging from R2adj = 0.9934 to 0.9963. When individual compliance components were considered (1/Ei, 1/Gij, or -v[ij]/Ei), correlation coefficients ranged from R2adj = 0.924 to 0.982. All four fabric measures performed equally well. It is concluded that volume fraction and fabric measures correlate highly with the apparent elastic properties of bone samples, provided that anisotropy and nonhomogeneity in the elastic properties of the trabecular tissue itself have negligible effects on the apparent properties. Whether this is true for bone in general remains to be seen, as only a subset was analyzed here. These methods, however, can be valuable in similar assessments of other subsets.
如果能直接从立体形态测量参数推断出松质骨的力学性能,那将是很有利的。为此,表观密度和平均截距长度作为骨量和方向性(结构)的度量,通常与在压缩试验中测定的骨样本的弹性特征相关。然而,尚未以这种方式建立完整和准确的关系。这可能不仅是由于立体形态测量和力学评估中出现伪影,还由于平均截距长度在表征骨结构的完整力学意义或立体形态测量中未考虑的小梁组织属性的非均质性方面存在固有不足。在本研究中,我们引入了一种计算机建模方法,可消除这些偏差。利用高分辨率三维计算机重建的松质骨标本进行立体形态测量,并用于微观结构有限元模型以模拟力学试验,无偏差的比较变得可行。目的是研究是否可以通过这种方式建立准确和完整的关系。考虑了四种不同的结构度量:平均截距长度和三种基于体积的度量。根据Cowin提出的数学关系,从结构张量计算顺应性矩阵,用于29个椎骨鲸骨标本。这些与直接从微观结构有限元模型模拟确定的顺应性常数相关。所有29个标本的九个正交各向异性弹性常数从其立体形态测量的结构和体积分数值得到了很好的预测,相关系数范围从R2adj = 0.9934到0.9963。当考虑单个顺应性分量(1/Ei、1/Gij或-v[ij]/Ei)时,相关系数范围从R2adj = 0.924到0.982。所有四种结构度量表现同样良好。得出的结论是,只要小梁组织本身弹性属性的各向异性和非均质性对表观属性的影响可忽略不计,体积分数和结构度量与骨样本的表观弹性属性高度相关。这里仅分析了一个子集,对于一般的骨是否如此仍有待观察。然而,这些方法在对其他子集的类似评估中可能很有价值。