The Structural Implications of a Unilateral Facial Skeletal Cleft: A Three-Dimensional Finite Element Model Approach

Linping Zhao, Ph.D., Jean E. Herman, R.N., M.S., M.B.A., Pravin K. Patel, M.D.
Objective: For children born with a unilateral facial skeletal cleft, oral motor function is impaired and skeletal development and growth are asymmetrical with regard to the midsagittal plane. This study was designed to verify that a unilateral skeletal cleft and its dimensions (i.e., depth and width) affect the severity of the asymmetric stress and strain distribution within the maxilla. Methods: A three-dimensional finite element model of a normal maxilla was developed from pediatric, subject-specific computerized tomography scan data. A clefting pattern then was introduced to simulate varying degrees of deformity in geometry, with the bone properties and boundary conditions held constant. The asymmetric index was introduced to quantify the asymmetrical stress and strain distribution within the maxilla with regard to the midsagittal plane. Results: The unilateral skeletal cleft led to a nonuniform, asymmetric stress and strain distribution within the maxilla: intensified on the noncleft side and weakened on the cleft side. As the depth of the unilateral cleft increased, the stress and strain distribution became increasingly asymmetric as measured by the asymmetric index. In contrast, the width of the cleft had minimal effect on the asymmetrical stress and strain distribution. Interpretation/Conclusion: These results implied that a child born with a unilateral cleft would be expected to have an asymmetric skeletal development between the noncleft and the cleft sides as a consequence of an asymmetric functional loading pattern. KEY WORDS: alveolar bone graft, biomechanics, cleft surgery, finite element analysis, midfacial skeletal growth, unilateral cleft palate

Clefts of the lip and palate are the most common craniofacial birth defect and the second most common congenital abnormality, with a birth prevalence ranging from 1 in 500 to 1 in 2000, depending on the population (Marazita and Mooney, 2004). Affected children may undergo multiple-staged reconstructive procedures to restore the underlying facial skeleton and overlying soft tissue envelope. In children with a unilateral cleft lip and palate (UCLP), abnormality of the facial skeleton in all three planes (transverse, sagittal, and coronal) is commonly present. Skeletal asymmetry is pronounced in the anterior part of the maxilla with a smaller maxillary width and mahel and Brejcha, 1983; S mahel height on the cleft side (S and Mullerova, 1986). There is a reduced vertical deDr. Zhao, Ms. Herman, and Dr. Patel are affiliated with Shriners Hospitals for Children, Chicago, Illinois. Dr. Patel also is affiliated with the University of Illinois at Chicago, Craniofacial Center, Chicago, Illinois, and Northwestern University, Feinberg School of Medicine, Chicago, Illinois. Submitted September 2006; Accepted April 2007. Address correspondence to: Dr. Zhao, Shriners Hospitals for Children, 2211 N Oak Park Avenue, Chicago, IL 60707. E-mail lzhao@ shrinenet.org. DOI: 10.1597/06-183.1 121

velopment of the maxilla on the cleft side (Mlsted and Dahl, 1990) as well as a collapse of the arch segments in the transverse plane (Aduss and Pruzansky, 1968; Berkowitz and Pruzansky, 1968). Clinically, it is observed that such skeletal asymmetry appears minor in infancy, as shown in Figure 1A, and becomes severe in adolescence, as shown in Figure 1B. Facial skeletal growth in a child born with a facial cleft is regulated by both intrinsic and functional factors such as genetics (Enlow, 1982), soft tissue matrix (Moss and Salentijn, 1969), and mechanical stimulation (McNeil, 1954) and is affected by iatrogenic causes as well (Ross, 1987). Among these, functional loads, specifically the occlusal forces, play an important role. As a mechanical stimulant, occlusal forces and corresponding stress and strain distribution fields within the skeletal components lead to strain-induced bone remodeling (Chierici, 1977; Carter et al., 1998; Frost, 2001, 2004; Mao and Nah, 2004). It is conceivable that the interaction between occlusal forces and the unilateral skeletal cleft leads to an asymmetrical stress and strain distribution in the maxilla and thus results in asymmetrical skeletal growth and development. Clinical patterns of unilateral facial clefts may range from minimal involvement of the primary alveolar segment

122 Cleft PalateCraniofacial Journal, March 2008, Vol. 45 No. 2

FIGURE 1 Asymmetric growth of the maxilla with UCP as demonstrated with the three-dimensional reconstructed CT scan of an infant (A) and an adolescent (B) in whom the cleft side is considerably smaller than the noncleft side.

of the anterior palate to complete involvement of the primary and secondary palates, as illustrated in Figure 2. Clinical observation suggests that the asymmetric growth of the maxilla with a unilateral cleft is related to the extent or the geometric dimension (i.e., depth and width) of the cleft (Fig. 1). However, it is not clear how the cleft pattern and its dimensions alter the stress and strain distribution within the maxilla and thus affect the facial skeletal growth and development. The aim of this study was to investigate, using a threedimensional linear elastic finite element analysis, the structural effects of a unilateral cleft and its geometric dimensions (i.e., depth and width) on the stress and strain distribution within the maxilla subjected to occlusal forces. This paper intends to verify the following hypotheses: (1) the unilateral facial skeletal cleft leads to a nonuniform, asymmetric stress and strain distribution within the maxilla during a functional task that can be quantified using an Asymmetric Index (AI), and (2) the geometric dimensions (i.e., depth and width) of the unilateral cleft affect the severity of the asymmetric stress and strain distribution within the maxilla and the midfacial skeleton.

METHODS Finite Element Model Design Subject-specific computerized tomography (CT) scans were obtained from a male adolescent with a normal maxilla following Institutional Review Board approval. Using ANALYZE AVW 4.0 (Biomedical Imaging Resource, Mayo Foundation, Rochester, MN), the CT scan data were reformatted and the voxels were resized using trilinear interpolation. After the hard tissue was separated from the soft tissue using a proper threshold value, the maxilla was separated from the mandible and the skull. A surface model of the maxilla was generated with triangular elements and then input into ABAQUS/CAE (ABAQUS Inc., Pawtucket, RI) for preprocessing. After manual editing (i.e., cleaning, repairing, and smoothing), volumetric meshes were generated using tetrahedral elements. The model consisted of 33,902 nodes and 143,084 tetrahedral elements with an edge length of 1.65 6 0.43 mm, which is smaller than the critical edge length of 2.11 mm. Such fine mesh ensured that the errors in calculated stress and strain

FIGURE 2 Typical patterns of a UCP on the maxilla: (A) missing a tooth and alveolar notch, (B) primary cleft palate, and (C) complete primary and secondary cleft palate.

Zhao et al., STRUCTURAL IMPLICATIONS OF UCP: A FEM APPROACH 123

FIGURE 3 Isometric (A) and submental view (B) of a FEM of a subject-specific maxilla (CM) to which a unilateral cleft was introduced. Pink and blue arrows indicate the fixed posterior and superior ends without displacement, whereas yellow arrows indicate the occlusal forces added on the teeth.

components would be less than 3.0%, based on a convergence test. This model served as the control model (CM), as shown in Figure 3. A clefting pattern then was introduced to the CM by removing corresponding volumetric elements. Four models were established to simulate various depths of the clinically observed unilateral cleft pattern: (1) absent second right incisor (MT); (2) alveolar ridge defect (AR), which modeled the primary cleft palate; (3) incomplete unilateral cleft palate (IUCP); and (4) complete unilateral cleft palate (CUCP), which models the complete primary and secondary cleft palate. Three additional models were used to simulate complete unilateral clefts, each of various widths: (1) absent 1.0 dental unit (CUCP); (2) absent 1.5 dental units (CUCPm); and (3) absent 2.0 dental units (CUCPw). For all finite element models (FEMs), the maxilla, simplified as the cortical bone, was modeled as a linear elastic body with an elastic modulus of 12.7 Giga Pascal (GPa) and a Poisson ratio of .3 (Hara et al., 1998; Tanaka et al., 2000). The use of this value was ensured by a sensitivity study. The sensitivity study was conducted using the CM at five levels of elastic modulus: 8.7, 10.7, 12.7, 14.7, and 16.7 GPa, and it showed that the alteration in the AI, as defined in the following section, was less than 4.13%. For all models, the inferior and posterior ends of the maxilla were fixed without movement while loads (nodal forces) were added to the teeth perpendicular to the occlusal plane, creating a total

force of 100 N. These loading and boundary conditions also are shown in Figure 3. Finite Element Analysis and the Asymmetry Index The finite element analysis was conducted using ABAQUS STANDARD V 6.4 (ABAQUS Inc., Pawtucket, RI). The computed stress and strain distribution pattern on the maxilla was studied in terms of von Mises stress, principal stresses, and principal strains. The von Mises stress is a scalar value that indicates the energy that has accumulated as a result of deformation. The values of the principal stresses and strains can be either positive or negative. A negative principal stress indicates that a compressive stress acted on the point along the corresponding principal direction and vice versa. A negative principal strain indicates that a shortened fiber acted along the corresponding principal direction at the point and vice versa. In order to quantify the asymmetrical stress and strain distribution with regard to the midfacial plane, the stress and strain components at selected anatomic sites were compared. These selected sites were paired between the cleft and noncleft maxillary halves: at the paranasal region and at the palatal alveolus of the canine teeth. Then, an AI was calculated accordingly. The AI was defined as a relative difference (absolute value) of a mechanical component (stress or strain) between two corresponding sites in a pair

124 Cleft PalateCraniofacial Journal, March 2008, Vol. 45 No. 2

and was calculated using the following formula:   2SN { SC   :100% AI ~  SN 0 z SC 0 

Here SN and SC were the values of the stress and strain component at the selected sites on either the noncleft or the cleft side, respectively. SN0 and SC0 were the values of the corresponding stress and strain component at the selected sites predicted by the CM. The error of the stress and strain component at a selected site, -Si, was estimated using the difference between the value of the stress and strain components at the selected site Si and that at its adjacent site (node or element), Sia: DSi ~ Si { Sia (i~N,C,N0,C0) 2

Considering the propagation of error (Meyer, 1975) and assuming that the errors of measured variables were independent, the AIs error, DAI, was estimated using the following formula: 2 S N 0 zS C 0 s DSN 0 2 zDSC 0 2 DSN 2 zDSC 2 zSN {SC 2 : SN 0 zSC 0 2 DAI ~

structures as well as of the error of the FEM. Similar von Mises stress distribution also was observed in the case of missing a tooth, as shown in Figure 4B. In the cases of maxilla with a unilateral cleft, the von Mises stress distribution was asymmetrical with regard to the midsagittal plane. It increased on the cleft side and decreased on the noncleft side, as shown in Figure 4C through 4E. The value of von Mises stress at stress concentration zones, for example, increased to 2.04 6 0.04 MPa on the noncleft side and decreased to 0.49 6 0.03 MPa on the noncleft side in the case of a complete unilateral cleft, as predicted by the CUCP model. A similar pattern also was observed in the principal stress and principal strain components. In the case of the incomplete unilateral cleft, as modeled by AR and IUCP models, a stress and strain concentration at the tip of the cleft was observed, which characterized a nonuniform stress and strain distribution. It should be noted that the AR model, which models the cleft alveolus, predicted the most intensive stress and strain concentration at the cleft tip. Effect of Cleft Depth Along with an increase in the depth of the unilateral cleft, modeled sequentially as the CM, MT, AR, IUCP, and CUCP models, the stress and strain components examined in this study intensified on the noncleft side but reduced on the cleft side, as demonstrated in Figure 4. For instance, the magnitude of von Mises and minimum principal stresses at the paranasal region, as shown in Figure 5A, increased on the noncleft side while decreasing on the cleft side along with an increase in the cleft depth, as shown in Figure 5B and 5C. The degree of asymmetrical distribution of these components, as quantified by the AI, increased accordingly, as shown in Figure 5D. The value of AI calculated from von Mises stress at the paranasal region was 10% 6 7% for the normal maxilla (CM) and 20% 6 8% for the case with a missing tooth (MT). This index, however, increased considerably to 68% 6 10% for the case with a cleft alveolus (AR), to 116% 6 9% for the primary cleft (IUCP), and to 144% 6 7% for the complete unilateral cleft (CUCP), as shown in Figure 5D. The similar relationship between the AI and the cleft depth also was obtained when the AI was calculated from other stress and strain components at all checked sites. Typical results are summarized in Table 1. It is to be noted that there was a slight difference between the values of the AIs calculated from different stress and strain components. The largest difference (16%) was between the AI computed from von Mises stress and that computed from the minimum principal stress. In summary, regardless which stress and strain components were used and which sites were studied, the overall trend of the AI as a function of the cleft depth was consistent: the larger the cleft depth, the higher the AI value.

RESULTS Stress and Strain Distribution Pattern on the Maxilla It was observed that the stress and strain distribution pattern on the normal maxilla was symmetrical with regard to the midsagittal plane, as predicted by the CM. On the contrary, the stress and strain distribution pattern on the maxilla with a unilateral cleft was asymmetrical with regard to the midsagittal plane: a higher level of stress and strain appeared on the noncleft side, as predicted by the models of the maxilla with a unilateral skeletal cleft (AR, IUCP, CUCP). As an aid to reviewing these simulated results, the frontal view of the maxilla with von Mises stress distribution is shown in Figure 4. On the normal maxilla, as seen in Figure 4A, von Mises stress distributed symmetrically with regard to the midsagittal plane. There was a pair of stress concentration zones at the nasomaxillary buttress next to the pyriform aperture on both the noncleft and the cleft side. The peak values of von Mises stress at these zones were up to 1.02 6 0.02 Mega Pascal (MPa) for the noncleft side and 1.12 6 0.07 MPa for the cleft side. Such a small difference between these peak values was indicative of the minor degree of asymmetry seen in all human facial skeletal

Zhao et al., STRUCTURAL IMPLICATIONS OF UCP: A FEM APPROACH 125

FIGURE 4 Frontal view of von Mises stress distribution pattern on (A) the normal maxilla (CM); (B) the maxilla with one tooth missing (MT); (C) the maxilla with unilateral cleft alveolus (AR); (D) the maxilla with an incomplete unilateral cleft (IUCP); and (E) the maxilla with a complete unilateral cleft (CUCP).

Effect of Cleft Width In combination with an increase in width of the complete unilateral cleft, stress and strain at the paranasal region remained almost the same on the noncleft side but decreased slightly on the cleft side. Consequently, there was a slight increase in the AI values. For example, at the palatal alveolus of the canine, as shown in Figure 6A, both von Mises stress and maximum principal strain held almost constant as the width of the cleft increased, as shown in Figure 6B and 6C, whereas the values of the AI slightly increased, as shown in Figure 6D. Compared with the error level, however, the increase in the AI as a function of the cleft width was trivial. A similar relationship between the AI and the cleft width also was predicted using the AIs that

were calculated from other stress and strain components at each examined site. Typical results are listed in Table 2. DISCUSSION Patients with facial clefts present with abnormal skeletal structure and facial morphology as a result of intrinsic, functional, and iatrogenic causes (Ross, 1987). In patients with a UCP, asymmetric development of the facial skeleton in all three planes (transverse, sagittal, and coronal) has been well documented in the literature (Aduss and mahel Pruzansky, 1968; Berkowitz and Pruzansky, 1968; S mahel and Mullerova, 1986; Mlsted and Brejcha, 1983; S and Dahl, 1990). There are a multitude of variables that influence the outcome leading to such asymmetry. One such

126 Cleft PalateCraniofacial Journal, March 2008, Vol. 45 No. 2

FIGURE 5 Asymmetrical stress distribution on the paranasal region (A) as a function of the cleft depth demonstrated by (B) the von Mises stress; (C) the minimum principal stress; and (D) the AI.

variable is the geometry of the cleft itself and the resulting differences in the transfer of functional loads that can influence subsequent upper facial skeletal development and remodeling. The unilateral cleft itself introduces an asymmetry in the facial skeleton, and the extent to which it contributes to the skeletal deformity remains to be elucidated. A biomechanical model, such as a FEM, would help to elucidate this outcome component. The finite element method previously had been used to model the load transfers within the skull (Tanne et al., 1989; Jinushi et al., 1997; Iseri et al., 1998; Gross et al., 2001) and the maxilla (Nagasao et al., 2005). However, none of these studies addressed the clinically relevant situation in which the facial skeleton is disrupted by a cleft. As the first step toward developing a clinically relevant, patient-specific

FEM, the present study focused solely on the biomechanical response to the geometrical dimensions of the unilateral cleft. For this basic study, we chose a normal, unaffected maxilla from which various cleft patterns were introduced artificially rather than using a real skull of an unrepaired unilateral cleft. This approach was chosen due to the significant asymmetry in all three planes, which varies from individual to individual as a consequence of many influential, uncontrolled factors, including skeletal remodeling itself. Analysis of how the geometry of the isolated cleft asymmetry influences the loading pattern in the facial skeleton will provide a basic reference point for future, more complex investigations. During masticatory function such as incising, masticating, and chewing, the functional loads are transmitted through the dentition to the maxilla and upper midfacial

Zhao et al., STRUCTURAL IMPLICATIONS OF UCP: A FEM APPROACH 127

TABLE 1

Asymmetry Index (AI) as a Function of the Depth of a Unilateral Facial Cleft (%)*

* CM 5 normal maxilla, control model; MT 5 missing teeth; AR 5 alveolar ridge defect; IUCP 5 incomplete unilateral cleft palate; CUCP 5 complete unilateral cleft palate.

skeleton. In an unaffected maxilla, the transmission of occlusal loads would be expected to result in symmetrical stress and strain distribution within the midfacial skeleton to the extent of the symmetry of the facial skeleton. The prediction of the normal maxilla model (CM) in the present study confirmed our belief. This model further revealed stress concentration zones at the medial nasomaxillary buttresses and at the lateral zygomatic-maxillary buttresses. These locations clinically correspond to the structural buttresses where the bone is thicker anatomically (Manson et al., 1980). In contrast to the normal maxilla, the occurrence of a unilateral skeletal cleft breaches the structural integrity of the skeleton and alters the transmission pattern of the functional loads within the maxilla, which is transmitted primarily through the noncleft side rather than the cleft side. Correspondingly, the stress and strain distribution on the maxilla with unilateral cleft is both nonuniform (a stress and strain concentration occurred at the tip of the cleft) and asymmetric with regard to the midsagittal plane (e.g., the stresses and strains intensified on the noncleft side and decreased on the cleft side), as predicted in our study using the FEMs of the maxilla with a unilateral cleft. These predictions verify our first hypothesis that the unilateral cleft leads to a nonuniform and asymmetric stress and strain distribution pattern on the affected maxilla. Such an asymmetric stress and strain distribution would suggest an asymmetrical skeletal development on the noncleft side versus the cleft side. Because a certain level of stress and strain, corresponding to the threshold strain, is required to trigger the strain-induced bone modeling (Carter et al., 1998; Frost, 2001, 2004; Mao and Nah, 2004), the skeletal segments with higher stress and strain levels, such as the noncleft side, would be expected to have a greater potential to develop and grow in response. In contrast, the segment with lower stress and strain levels,

such as the cleft side, is less likely to further develop and grow. This may provide a possible explanation, from a biomechanical point of view, as to why the individual with a UCP develops a differential between the two maxillary segments both in volumetric size and in spatial dimensions, as shown in Figure 1, which has been documented in previous studies (Mlsted and Dahl, 1990; Peltomaki et al., 2001). However, our prediction does not exclude the effect of other intrinsic, extrinsic, and iatrogenic factors that may contribute to the midfacial skeletal asymmetry. The present study predicted that the geometrical dimensions of a unilateral cleft affect the asymmetrical stress and strain distribution as well. Along with the increase in the cleft depth, stress and strain increased on the noncleft side and decreased on the cleft side, as we demonstrated using the FEMs. Correspondingly, the degree of asymmetry in stress and strain distribution, as indicated by the AI, considerably increased. Along with the increase in the cleft width, though, the increase of the AI was not considerable within the scope of our current study. This supports that the width of the alveolar cleft in patients with UCLP is not related to their subsequent facial growth (Schwartz et al., 1984; Suzuki et al., 1993; Johnson et al., 2000). These results verify our second hypothesis that the dimensions of the unilateral cleft considerably affect the degree of asymmetry in the stress and strain distribution on the maxilla. Moreover, these predictions suggested that the effect of the cleft depth is more significant than the cleft width on the degree of asymmetry in the stress and strain distribution on the maxilla. As a preliminary study, certain assumptions were necessary to simplify the model and to emphasize the structural analysis of the midfacial skeleton. The application of a linear elastic material model to the maxilla neglected the nonlinear dynamic responses of the bony

128 Cleft PalateCraniofacial Journal, March 2008, Vol. 45 No. 2

FIGURE 6 Asymmetrical stress and strain distribution on the palatal alveolus of the canine (A) as a function of the cleft width demonstrated by (B) the von Mises stress; (C) the maximum principal strain; and (D) the AI.

materials. The homogeneous and isotropic assumption simplified the structure of the maxillary bone and especially of the unique load-bearing feature of the dentition and surrounding soft tissues such as the periodontal ligaments. The variations in the prediction of the stress and strain distribution and the AI caused by these assumptions were addressed in our study via a sensitivity study. The sensitivity study of the elastic modulus from 8.7 to 16.7 GPa covered a wide range of possible values for human maxilla. The maximum difference in the AI predictions was 4.13%, which was less than the root-mean-square errors of the AI. This ensured that the simplified material model used in this study had a minimal impact on the prediction

of the AI. Even so, the interpretation of such model predictions is limited to the relative lower loading range, where the elastic deformation of the skeletal structure is dominant. Additionally, assumptions about material properties of the dentition and periodontal tissues, together with the loading condition at the dentition, would lead to less accurate predictions within the immediate dental regions. Whereas the material model was more accurate regarding the body of the maxilla and upper facial skeletal regions, the loading effect, according to St. Venants principle, would be less likely to significantly influence the predictions in the regions of interest that are away from where the

Zhao et al., STRUCTURAL IMPLICATIONS OF UCP: A FEM APPROACH 129

TABLE 2

Asymmetry Index (AI) as a Function of the Width of a Unilateral Facial Cleft (%)*
CM CUCP 1 Dental Unit CUCPm 1.5 Dental Unit CUCPw 2 Dental Unit

Stress and Strain, Components and Sites

0 Dental Unit

von Mises stress, paranasal region Minimum principal stress, paranasal region von Mises stress, palatal alveolus Maximum principal strain, palatal alveolus

10 24 9 22

6 6 6 6

7 7 8 3

144 128 154 194

6 6 6 6

7 6 14 5

151 133 158 199

6 6 6 6

7 7 14 5

159 139 160 203

6 6 6 6

7 8 14 5

* CM 5 normal maxilla, control model; CUCP 5 complete unilateral cleft palate of 1 dental unit in width; CUCPm 5 complete unilateral cleft palate of 1.5 dental units in width; CUCPw 5 complete unilateral cleft palate of 2 dental units in width.

loads were applied. Therefore, the stress and strain pattern on the body of the maxilla away from the dentition would be expected to be more accurate. Finally, our choice of superimposing an artificial cleft in the normal maxilla instead of deriving a FEM from a patient-specific skull with an untreated unilateral cleft is a significant oversimplification of the cleft skeletal deformity that occurs in nature. Though artificial, our model allowed us to eliminate any skeletal asymmetry due to asymmetric growth and development caused by uncontrolled variables inherent in a FEM derived from a cleft patient. Our model allowed us to artificially eliminate abnormal skeletal development as a result of bone remodeling and dental movement that depends upon intrinsic, functional, and extrinsic (treatment protocol) factors that vary from individual to individual. Such simplification was necessary to focus on the investigation of the biomechanical effects of the clefts as a structural defect alone. However, the results predicted in the present study should not be overinterpreted without clinical correlation. Correlation with similar models derived from patients with clefts (patient-specific FEMs) is the subject of future studies. CONCLUSION The presence of a unilateral cleft leads to nonuniform, asymmetrical stress and strain distribution within the maxilla and midfacial skeleton. The dimensions of the cleft influence the asymmetric distribution of the stress-strain pattern. Increased depth of the unilateral cleft leads to an increasingly greater asymmetric stress and strain distribution within the maxilla and midfacial skeleton. In contrast, the width of the unilateral cleft has considerably less effect than the depth of the cleft on the asymmetrical stress and strain distribution. The asymmetric development of the midfacial skeleton frequently seen in children with a unilateral cleft might be attributed in part to an asymmetric stress and strain pattern subjected to the functional loads simply as a result of the structural defect.
Acknowledgments. This research was supported by Shriners Hospitals for Children. The authors would like to acknowledge Mr. Kurt Peterson for his illustrations and Miss Laura Mueller for proofreading the manuscript.

REFERENCES
Aduss H, Pruzansky S. Width of cleft at level of the tuberosities in complete unilateral cleft lip and palate. Plast Reconstr Surg. 1968;41:113123. Berkowitz S, Pruzansky S. Stereophotogrammetry of serial casts of cleft palate. Angle Orthod. 1968;38:136149. Carter DR, Beaupre GS, Giori NJ, Helms JA. Mechanobiology of skeletal regeneration. Clin Orthop Relat Res. 1998:S41S55. Chierici G. Experiments on the influence of oriented stress on bone formation replacing bone grafts. Cleft Palate J. 1977;14:114123. Enlow DH. Handbook of Facial Growth. Philadelphia: WB Saunders; 1982. Frost HM. From Wolffs law to the Utah paradigm insights about bone physiology and its clinical applications. Anat Rec. 2001;262:398419. Frost HM. A 2003 update of bone physiology and Wolffs law for clinicians. Angle Orthod. 2004;74:315. Gross M, Arbel G, Hershkovitz I. Three-dimensional finite element analysis of the facial skeleton on simulated occlusal loading. J Oral Rehab. 2001;28:684694. Hara T, Takizawa M, Sato T, Ide Y. Mechanical properties of buccal compact bone of the mandibular ramus in human adults and children: relationship of the elastic modulus to the direction of the osteon and the porosity ratio. Bull Tokyo Dent Coll. 1998;39:4755. Iseri H, Tekkaya A, Oztan O, Bilgic S. Biomechanical effects of rapid maxillary expansion on the craniofacial skeleton, studied by the finite element method. Eur J Orthod. 1998;20:347356. Jinushi H, Suzuki T, Naruse T, Isshiki Y. A dynamic study of the effect on the maxillofacial complex of the face bow: analysis by a threedimensional finite element method. Bull Tokyo Dent Coll. 1997;38:3341. Johnson N, Williams A, Singer S, Southhall P, Sandy J. Initial cleft size does not correlate with outcome in unilateral cleft and palate. Eur J Orthod. 2000;22:93100. Manson PN, Hoopes JE, Su CT. Structural pillars of the facial skeleton. Plast Reconstr Surg. 1980;66:5462. Mao JJ, Nah HD. Growth and development hereditary and mechanical modulations. Am J Orthod Dentofac Orthop. 2004;125:676689. Marazita ML, Mooney MP. Current concepts in the embryology and genetics of cleft lip and cleft palate. Clin Plast Surg. 2004;31:125140. McNeil C. Oral and Facial Deformity. London: Sir Isaac Pitman and Sons; 1954. Meyer S. Propagation of error and least squares. In: Meyer S, ed. Data Analysis for Scientists and Engineers. New York: Wiley; 1975:3948. Mlsted K, Dahl E. Asymmetry of the maxilla in children with complete unilateral cleft lip and palate. Cleft Palate J. 1990;27:184190. Moss ML, Salentijn L. The primary role of functional matrices in facial growth. Am J Orthod. 1969;55:566577. Nagasao T, Nakajima T, Kimura A, Kaneko T, Jin H, Tamaki T. The dynamic role of buttress reconstruction after maxillectomy. Plast Reconstr Surg. 2005;115:13281340. Peltomaki T, Vendittelli BL, Grayson BH, Cutting CB, Brecht LE. Associations between severity of clefting and maxillary growth in

130 Cleft PalateCraniofacial Journal, March 2008, Vol. 45 No. 2

patients with unilateral cleft lip and palate treated with infant orthopedics. Cleft Palate Craniofac J. 2001;38:582586. Ross R. Treatment variables affecting growth in unilateral cleft lip and palate. Cleft Palate Craniofac J. 1987;24:577. Schwartz BH, Long RE Jr, Smith RJ, Gipe DP. Early prediction of posterior crossbite in the complete unilateral cleft lip and palate. Cleft Palate Craniofac J. 1984;21:7681. mahel Z, Brejcha M. Differences in craniofacial morphology between S complete and incomplete unilateral cleft lip and palate in adults. Cleft Palate J. 1983;20:113127. mahel Z, Mullerova Z. Craniofacial morphology in unilateral cleft lip S and palate prior to palatoplasty. Cleft Palate J. 1986;23:225232.

Suzuki A, Mukai Y, Ohishi M, Miyanoshita Y, Tashiro H. Relationship between cleft severity and dentocraniofacial morphology in Japanese subjects with isolated cleft palate and complete unilateral cleft lip and palate. Cleft Palate Craniofac J. 1993;30:175181. Tanaka E, Rodrigo DP, Miyawaki Y, Lee K, Yamaguchi K, Tanne K. Stress distribution in the temporomandibular joint affected by anterior disk displacement: A three-dimensional analytic approach with finite element method. J Oral Rehab. 2000;27:754759. Tanne K, Hiraga J, Kakiuchi K, Yamagata Y, Sakuda M. Biomechanical effect of anteriorly directed extraoral forces on the craniofacial complex: a study using finite element method. Am J Phys Anthropol. 1989;11:95.