750 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA J Oral Maxillofac Surg 52:750-751,1994 DISCUSSION zyxwvutsrqpo Discussion Biomechanical Rationale for Surgical- Orthodontic Expansion of the Adult Maxilla Paul C. Dechow, PhD Baylor College oJ ‘Dentistr.v. Dallas. TX Analyses of loading patterns in the human craniofacial skeleton are greatly hindered by both the complexity of the structures and the inaccessibility of many components in vivo. The result is a long tradition of modeling approaches to hu- man craniofacial form and biomechanics. Modeling ap- proaches include 1) the use of various experimental animals, such as primates, pigs, or rats, depending on the nature of the investigation, 2) mathematical modeling techniques. ranging from simple constructs based on beam theory to more refined finite element models of mandibular form,’ and (3) the use of analogues of craniofacial structure. The most com- monly used structural analogues are the actual skeletal com- ponents. These remains must be considered analogues of themselves in their in vivo state because the preparation of the specimens, which often includes the removal of soft tissues and drying, results in alteration of the biomechanical prop- erties. The study by Shetty and colleagues takes another ap- proach, the construction of a photoelastic analogue of the human skull. The construction of such a model is a problem- atic task. but its rigors are sometimes compensated for by a unique property. The photoelastic mode1 allows direct vi- sualization of the state of stress in the structure, if appropriate procedures are followed.’ The common requirement of all models is the assessment of reliability and validity, and, as in the early stages of many models, the photoelastic skull model requires further testing for optimal interpretation of results. Despite this problem. the work by Shetty et al provides some interesting insights into both the normal and experimental loading of the facial skeleton. This discussion will address several issues: the re- liability and validity of the model, and its implications for understanding facial biomechanics. Reliability Reliability is the internal consistency and replicability of a model. Reliability also concerns the repeatability of the experiments conducted with a model. In general, internal consistency is not a major concern in a carefully constructed photoelastic model. Shetty et al have noted the steps they have taken in constructing the model, including assessment of important features such as a lack of residual stresses in the unloaded model. One problem in reliability, variation in stresses following loading, could serve as the basis for further research. The results are from testing with a single maxillary expansion appliance. However. inspection of these appliances suggests that variations in the placement of the arms and the central screw, and in the fit of the appliance on the model, has an impact on the relative amounts of lateral force anteriorly and posteriorly. Further experiments could test the sensitivity of loading patterns in the facial skeleton to variations in appli- ance design or construction. It would be interesting to see what changes in appliance design would be necessary to shift the greatest stresses in the hard palate from the posterior edge to a more anterior position. Another problem in reliability is the potential for sequence error in the experimental design. Would similar conclusions be reached about the relative importance of the various os- teotomies in enabling maxillary expansion if the sequence of osteotomies was altered? This problem also points to the in- direct method of assessing forces required for separation of the midpalatal suture and poses another area for research. The conclusions about the decreased forces required for maxillary expansion are suggested indirectly by greater stresses at sites distant from the maxilla following the various oste- otomies. A different approach, such as a modified appliance with an internal load cell or outfitted with strain gages, could give direct measurements of expansion forces. These forces could be correlated with stress magnitudes. Validity Validity requires measurement of the similarity between the model and the biologic structure. Otherwise, we cannot conclude the extent to which we can generalize to living tis- sues. Principles applied to the validity of the mathematical representation of form3 can also be used to assess structural analogues. To address the question of accuracy, the robustness of the model must be determined, ie, how much do changes in the features of the model affect the measures of outcome? In a photoelastic model of the cranium, there are a large number of differences with living tissue. Such differences, to name a few. include a lack of soft tissues, variations in ma- terial properties, differences in the sutural and periodontal properties, and variations in the internal structures of bones. We must ask, are these differences significant given the nature of the investigation? The most troubling difference is the variation in the internal structures of bones compared to the homogeneous structure of the model. On the gross level, this problem is most apparent in the representation of internal spaces in models of cranio- facial bones. For instance, an external cast of the maxilla or the ethmoid does not provide internal representation of the maxillary or ethmoid sinuses. Yet a solid maxilla or ethmoid in a photoelastic mode1 of the human face is likely to create large differences with actual stresses in real bone. Such dif- ferences, especially in the maxilla, are also likelv to change the effectiveness of a simulated zygomaticomaxillary oste- otomy in relieving local stresses. A similar argument can be made for the artificial sutures in the photoelastic model. In the model, the sutures tended to be sites of stress concentration. In a real cranium. the states of stress around sutures are much less clear and prob- ably vary depending on the suture and its location. In open sutures, complex patterns of connective tissue and ligaments may relieve stress rather than concentrate it. In some closed sutures, there may actually be little difference in behavior from that of the surrounding bone. Little empirical or ex- perimental work on sutural biomechanics is available to test the above conjectures, although the mechanical properties of crania1 sutures have been shown to change with increased