Cervical spine models for biomechanical research.

Biomechanical models have been used for the understanding of the basic normal function and dysfunction of the cervical spine and for testing implants and devices. Biomechanical models can be broadly categorized into four groups: 1) Physical models, made of nonanatomic material (e.g., plastic blocks), are often used for the testing of spinal instrumentation when only the device is to be evaluated. 2) In vitro models consisting of a cadaveric spine specimen are useful in providing basic understanding of the functioning of the spine. Human specimens are more suitable for these models than are animal specimens whenever anatomy, size (for instrumentation), and kinematics are important. Animal specimens are less costly, easier to obtain, and often have less variability but should be used with care because of the absence of anatomic fidelity with the human. 3) In vivo animal models provide the means to model living phenomena, such as fusion, development of disc degeneration, instability, and adaptive responses in segments adjacent to spinal instrumentation. Choosing the appropriate animal is important. The appropriate animal should have spinal loading, kinematics, kinetics, vertebral size, and healing-fusion rates as similar to those in humans as possible. For better interpretation of in vivo animal experimental results, in vitro biomechanical study using the same animal cadaveric specimen is useful but has not been used routinely. 4) Computer models are developed from mathematical equations that incorporate geometry and physical characteristics of the human spine and may be advantageously used for problems that are difficult to model by other means. Examples are the changes in disc and vertebral stresses in response to graded transection of facet joints and the study of changes in endplate loading caused by disc degeneration. Because these models are purely mathematical, their validation is essential. Validation is best achieved by first incorporating high-quality geometry and physical characteristics of the human spine and then comparing the model predictions with experimental observations. Sometimes an enthusiastic researcher may use a computer model beyond its validation boundary, making the model's predictions unreliable. In general, it is important to remember that a biomechanical model, similar to any other model, represents only a certain aspect of the real living human being. The aspect chosen for representation should be selected with great care. The model should be designed to answer specifically the question asked. Its predictions are valid only within the boundaries of assumptions and limitations that it incorporates.