Center for Orthopaedic Biomechanics, University of Denver, Denver, CO 80208, USA.
Med Eng Phys. 2013 Oct;35(10):1450-6. doi: 10.1016/j.medengphy.2013.03.021. Epub 2013 May 3.
By characterizing anatomical differences in size and shape between subjects, statistical shape models enable population-based evaluations in biomechanics. Statistical models have largely focused on individual bones with application to implant sizing, bone fracture and osteoarthritis; however, in joint mechanics applications, the statistical models must consider the geometry of multiple structures of a joint and their relative position. Accordingly, the objectives of this study were to develop a statistical shape and alignment modeling (SSAM) approach to characterize the intersubject variability in bone morphology and alignment for the structures of the knee, to demonstrate the statistical model's ability to describe variability in a training set and to generate realistic instances for use in finite element evaluation of joint mechanics. The statistical model included representations of the bone and cartilage for the femur, tibia and patella from magnetic resonance images and relative alignment of the structures at a known, loaded position in an experimental knee simulator for a training set of 20 specimens. The statistical model described relationships or modes of variation in shape and relative alignment of the knee structures. By generating new 'virtual subjects' with physiologically realistic knee anatomy, the modeling approach can efficiently perform investigations into joint mechanics and implant design which benefit from population-based considerations.
通过对不同个体间解剖结构大小和形状的差异进行特征描述，统计形状模型可以实现基于人群的生物力学评估。统计模型主要集中于对个体骨骼的研究，应用于植入物尺寸、骨骼断裂和骨关节炎；然而，在关节力学应用中，统计模型必须考虑关节多个结构的几何形状及其相对位置。因此，本研究的目的是开发一种统计形状和对齐建模（SSAM）方法，以描述膝关节结构的骨骼形态和对齐的个体间变异性，展示统计模型在训练集中描述变异性的能力，并生成用于有限元关节力学评估的真实实例。统计模型包括来自磁共振成像的股骨、胫骨和髌骨的骨骼和软骨表示，以及在实验膝关节模拟器中已知加载位置的结构的相对对齐，该模型使用 20 个样本的训练集。统计模型描述了膝关节结构的形状和相对对齐的关系或变化模式。通过生成具有生理上真实的膝关节解剖结构的新“虚拟个体”，该建模方法可以高效地进行关节力学和植入物设计研究，从而受益于基于人群的考虑因素。
