tissue analogs since this process requires testing the target tissues and analogs under the same conditions. Finally, cadavers available for such testing are generally drawn from a pool consisting of the elderly and diseased - which will ultimately yield suspect data. For these reasons and more, we do not consider synthetic tissue analogs based on cadaver or literature data to be effective replacements for live tissue.

Testing Living Tissue
Now that the targeted properties list has been finalized and data source has been selected the testing begins. The simple femoral model discussed previously involves at least two components (artery and support tissue) made from two different tissue analogs. We will assume for the sake of discussion that this model will be used primarily to evaluate abrasive tissue damage and the ease of tracking through the artery. If it is further assumed that the analog materials will be designed around porcine tissue properties then the animal must be sourced to collect the required samples for testing. It is important to remember here that tissue begins to decompose immediately after death so preserved samples cannot be used. It is best practice to collect data from tissues in situ using specialized instruments. However, when this is not possible the samples must be surgically removed (minimizing trauma), fixtured in a heated blood bath, and tested immediately after harvest to keep the sample alive during testing. In the current example the tests performed would include intimal abrasion resistance and coefficient of dynamic friction, as well as shear strength and penetration resistance of the vessel wall. However, other properties might be included as well. Of course, the desired end product is a set of analog materials that mimic the physical properties of the target tissue, so after the analogs are formulated their performance will be validated by performing the same tests under identical conditions.

Part of the design process involves prioritizing the various elements on the targeted properties list. Less important properties should be placed further down the list and given a lower priority during the tissue design process. This is critical because design criteria become progressively more difficult to satisfy as the list of target properties grows. As a practical matter the list should be limited to three or fewer properties if at all possible. In fact, if more complex model behavior is required than this restriction will allow then the number of tissue analogs for the component should be increased instead. For example, a synthetic rectus femoris muscle could be more easily constructed from three two-target tissue analogs than one six-target analog. In addition, any organ comprised of several analogs will exhibit a more complex and realistic response than one constructed from a single (multi-property) analog. The downside of this approach is that model fabrication and design costs are increased, but the improvement in performance may be substantial. In the case of most complex artery models, the arterial wall would probably be composed of at least three different (intima, media, and adventitia) tissue analogs, and the remainder of the model might employ multi-part components for skin, fat, muscle, and other tissues.

Synthetic Tissue Design
The composition of the materials used to construct individual anatomical structures is unimportant as long as the relevant properties are accurately modeled. However, typical engineering materials (engineering plastics, natural rubbers, organosilicates, etc) are inadequate for such applications, and in cases where soft tissues are modeled it will generally be advantageous to employ properly designed synthetic tissues. Such materials are designed to mimic a specific tissue (muscle, tendon, intima, etc) so that the resulting synthetic body part will contain appropriate levels of water, fiber,