To determine if density measurements of several maxillary regions in Hounsfield Units (HU) and outcomes of rapid maxillary expansion (RME) are correlated. Is correlation powerful enough to give us direct information about maxillary resistance to RME?
Twenty-two computed tomographic (CT) scans (14 years) are used in this archive study. Two CT records were collected, one before RME (T1) and one after 3 months of retention period (T2). Maxillary measurements were made using dental and skeletal landmarks in first molar and first premolar slides to measure the effects of RME. Density of midpalatal suture (MPSD) and segments of maxillary bone is measured in HU at T1. Correlation analysis was conducted between density measurements and maxillary variables. Regression analysis was then performed for variables that showed positive correlation.
There was no correlation between density and skeletal measurements. Intermolar angle (ImA) in molar slice showed statistically significant correlation with density measurements. The ImA variable showed the highest correlation with MPSD in frontal section (r = 0.669, P < .01).
There is correlation of 32.1–43.3% between density measurements and ImA increase. Our density measurements explain a certain percentage of ImA increase, but density is not the only and definitive indicator of changes after RME.
Rapid maxillary expansion (RME) is a widely used orthopedic procedure designed to increase maxillary transverse dimension in young patients by midpalatal and circumferential sutural opening. Maxillary expansion is recommended in certain clinical situations unless there is a systemic contraindication. By contrast, the decision to use surgically assisted RME (SARME) is not as easy, especially in young adults. With increasing age, the resistance of the maxilla to expansion increases as a result of three major causes: midpalatal synostosis, midpalatal interlocking, and circummaxillary rigidity.1 These important factors are clinically unmeasurable.
There are several proposed techniques to assist clinicians in choosing surgical assistance, including skeletal maturity assessment by hand-wrist radiography or cervical vertebra maturation2,3 and midpalatal suture assessment before and after RME by occlusal radiography.4 However, neither of these methods serves as a definitive indicator for SARME, and more importantly, these methods do not provide a direct evaluation of the maxillofacial area.
In orthodontics, three-dimensional (3D) imaging is indicated for numerous situations, such as the evaluation of impacted teeth,5 bone graft evaluation in cleft regions,6 alveolar bone analysis prior to placing temporary orthodontic anchorage devices,7 and assessment of the impact of RME on nasomaxillary structures.8 Computed tomography (CT) allows mineral density quantification in specific jaw sites in Hounsfield Units (HU) based on grayscale differences. This property of CT has been widely used, especially in the field of implantology, and is applicable to orthodontics.
The present study assesses the relationship between the CT density of several maxillary regions and the RME outcomes and determines the diagnostic utility of CT density in predicting maxillary resistance to RME. CT may potentially prove beneficial for clinicians choosing corticotomy, and it may provide measurable diagnostic information related to the maxillofacial area. SARME is a time- and energy-intensive procedure for the clinician and an invasive and costly procedure for the patient; therefore, the ability to thoroughly assess the need for SARME will benefit both.
MATERIALS AND METHODS
The parents of each patient provided informed consent to participate in the study, which was approved by the ethical committee of Marmara University, Institute of Health Sciences. Archived CT scans of 22 patients (11 males, 11 females; mean age: 14.25 years) at the Marmara University Dental Faculty, Department of Orthodontics, were evaluated retrospectively. The inclusion criteria were as follows: patients must be aged 11–17 years and have maxillary constriction with bilateral posterior crossbite, permanent dentition, no systemic disease, no previous orthodontic treatment, and no periodontal disease history.
RME was performed in each patient as a part of the orthodontic treatment comprising an acrylic–cap splint-type hyrax appliance, rigid bonded expansion appliance with a hyrax screw in the palate, and an acrylic cap splint covering the palatal and occlusal premolar and molar surfaces. The appliance was bonded using dual-cure glass-ionomer cement. Patients were instructed to activate the screw twice daily. Expansion was continued until the upper first molar palatal cusp tips touched the lower first molar buccal cusps; the amount of expansion was monitored at weekly examinations. At the end of the active expansion phase, the hyrax screw was stabilized with a stainless-steel ligature wire and flowable composite. The appliance remained in the mouth for 3 months to ensure retention. Afterwards, the expansion appliance was removed, and a transpalatal arch (TPA) with arms extending from the molar bands to the premolars was immediately cemented to further ensure retention.
Two CT records were collected from each patient, one before RME (T1) and one after the 3-month retention period (T2). The second CT was performed after removing the acrylic cap splint and before placing the TPA to prevent metallic artifacts that can influence imaging quality and diagnostic accuracy. CTs were collected at the same hospital using a spiral CT machine (Siemens Sensation 40, Siemens Medical Solutions of Siemens, Erlangen, Germany) at 120 kV, 80 mAs, with a 12.6 × 12.6-cm field of view and a 512 × 512-pixel matrix. Axial 0.3-mm increment slices comprising the entire cranium were obtained. Dicom CT data were analyzed with 3D prototyping software (Mimics v.16.0, Materialise, Leuven, Belgium). A specialized pillow and two perpendicular laser light beams stabilized and standardized the head position during scanning. The horizontal beam was adjusted parallel to the Frankfurt horizontal plane, and the vertical beam passed through the midsagittal plane.
The RME maxillary effects were measured using dental and skeletal landmarks (Figures 1 and 2) and compared between pre- and postexpansion coronal images. Dental and skeletal measurements were obtained from the coronal images passing through the center of the maxillary first molar and premolar palatal root apices. Bone density was measured only on pre-expansion CTs.
The following parameters were measured: apical width (AW)—the distance between the maxillary first molar palatal root apices; crown width (CW)—the distance between the maxillary first molar mesiobuccal cusp tips; and intermolar angle (ImA)—the angle between the maxillary first molar axes (molar axis defined as the line connecting the palatal root apex and mesiobuccal cusp tip of the maxillary first molar) (Figure 1).
The following skeletal parameters were measured: anterior nasal spine width (ANSW)—the distance between the right and left anterior nasal spine points measured on 3D skull images (Figure 2); alveolar width (AlvW)—the width between the most coronal points of the maxillary first molar buccal alveolar crest; and maxillary width (MaxW)—the width between the maxillary buccal cortexes of the maxillary first molars, tangent to the hard palate.
Midpalatal suture density on axial cut
On axial slice passing through the ANS and midpalatal suture, the center of the incisive foramen was marked as “O.” Points at 5 mm anterior (point A) and 5, 10, and 15 mm posterior (P1, P2, and P3, respectively) to “O” were also marked. Then, 1-mm2 squares were drawn at A, P1, P2, and P3. The square area densities were measured, averaged, and recorded as the midpalatal suture density on axial cut (MPSD axial) (Figure 3).
A 2-mm-wide rectangle was drawn on the midpalatal suture on the frontal view, limited by the nasal floor cortex superiorly and the palatal vault cortex inferiorly and centered on the midsagittal reference line. The rectangle density was measured at points A, P1, P2, and P3, and the average was recorded as the MPSD on frontal cut (MPSD frontal) (Figure 4).
3D skull images were obtained in the following planes: lateral, two planes (right and left) parallel to the midsagittal plane and passing through the most superior point to the frontomaxillary sutures; inferior, the plane passing through the ANS and the most proximal points of the right and left pterygoid hamulus (designated as the ANS-Sphenoid plane); posterior, the plane perpendicular to the ANS-Sphenoid plane and passing through the most proximal points of the right and left medial sphenoid bone plates; and superior, the plane parallel to the ANS-Sphenoid plane and passing through the right infraorbital foramen.
The resultant bone model was divided into three parts, anterior, middle, and posterior, by two vertical planes passing through the greater palatinal foramen and zygomatic process (Figure 5). Dental structures remaining in the models were erased from the image, and the bone density was recorded.
Midpalatal suture in 3D
The 3D skull reconstruction was divided by two planes parallel to the midsagittal plane and passing 3 mm lateral to the ANS bilaterally (Figure 6). Vomer and dental structures were erased, and the bone density was recorded.
The SPSS for Windows 15.0 (SPSS, Chicago, Ill) program was used for statistical analyses. Descriptive statistical methods (mean, standard deviation [SD]) and pre-/postexpansion quantitative data were compared using the paired samples t-test. Normally distributed data were analyzed with the Pearson correlation analysis; data not normally distributed were analyzed with the Spearman's rho correlation analysis. A simple regression analysis assessed the relationship between bone density and the ImA increase. Data reliability was assessed by the intraclass correlation coefficient (ICC). Results were evaluated at a P < .05 significance level and a 95% confidence interval.
The bone density minimum, maximum, mean, and SD are summarized in Table 1. Table 2 summarizes the mean, SD, and significance values of pre- and postexpansion dental and skeletal measurements. There were significant differences in the pre- and postexpansion values of all of the variables evaluated (P < .01). Density, skeletal, and dental measurements were also evaluated according to gender and age (mean age: 14 years; group 1, <14 years; group 2, >14 years). There were no statistically significant differences (P > .05) between male and female subjects or between subjects <14 and >14 years of age.
The correlation analyses for all of the variables are shown in Tables 3 and 4. The ImA molar slice showed a statistically significant correlation with all density measurements and was the most correlated with the MPSD frontal. There was a significant (P < .01) 66.9% correlation between the ImA increase and the MPSD frontal value. To ascertain the power of each density measurement on the ImA increase a regression analysis was performed for the ImA (molar) and density measurements. Correlation between bone density and other variables was not consistent, and therefore a regression analysis was not performed.
The regression analysis between the MPSD axial and ImA (Table 5) yielded an R2 value of 0.321, indicating that 32.1% of the ImA change can be attributed to the MPSD axial value (Table 5A). This model is statistically significant (P = .006; P < .01) (Table 5B). Table 5C shows the estimated coefficients and t values. A 1-unit increase in the MPSD axial value caused a 0.036× ImA increase. The t value for this coefficient was statistically significant. The final model is represented by the following equation:
The R2 values corresponding to each density measurement and resulting equation model are summarized in Table 6. The MPSD 3D had the greatest effect on the ImA, with an R2 = 43.3%.
All measurements were repeated by the same observer 3 months later. The ICC was close to 1.00. The method error evaluation is shown in Table 7. Density measurements and maxillary measurements were reliable and reproducible with an insignificant error and without affecting the results.
In current orthodontic dentistry, there are protocols guiding clinicians to choose between RME and SARME. However, studies investigating suture maturation have offered conflicting interpretations of the clinical principles used to evaluate skeletal age radiographically. A study by Schlegel et al.9 found that the youngest donor with ossification was 23 years old, yet a 50-year-old patient did not have ossification. Persson and Thilander10 reported that palatal sutures may become obliterated during the juvenile period, but a marked closure was rare prior to the third decade of life. They found the earliest closure in a 15-year-old girl, and bony spicules appear to have been observed between the ages of 15 and 19 years. Notably, only 5% of sutures are closed by 25 years, and 5% of suture closures with mechanical interlocking and circummaxillary rigidity can be broken without a corticotomy.1 In a more recent microcomputed tomography study, Korbmacher et al.11 concluded that the suture obliteration index was generally low, varied between individuals, and did not correlate with chronological age. Interdigitation was also independent of age. Collectively, these histological results indicate that there is a wide range for the corticotomy upper age limit, which is especially critical between the ages of 15 and 25 years. At present, the potential benefit from RME cannot be firmly predicted even in patients over 15 years of age prior to attempting the expansion clinically.
CT is widely used in implant dentistry because it offers a 3D analysis that enables mineral density quantification at specific jaw sites in HU.12 Several comparative studies confirmed the reliability and high accuracy of 3D CT for quantitative and qualitative analyses13–15 and concluded that CT is a valuable diagnostic supplement to subjective bone density evaluation.16–18 Cone beam computed tomography (CBCT) is preferred in dentistry as a result of the lower radiation dose it affords compared to that required in standard CT. However, standardization between CBCT machines is low,7 which causes Hounsfield scale variability. Therefore, despite the lower radiation dose, CBCT is unreliable for comparing and measuring the HU. Although CT is not in routine use either for orthodontics or research, its suitability for measuring density makes it the modality of choice in this study. Regardless, the decision to perform CT should be limited to cases in which low-dose CBCT is inadequate.
Head orientation was poor in some subjects, despite precautions to stabilize it. This can be overcome during distance and angular measurements by using well-defined anatomic references. However, poor cranium orientation particularly affected the suture density measurements on sectional views because the suture is a fine line and requires more precise measurement. Digital reorientation can be performed, but this maneuver decreases image quality and was thus avoided. Images were evaluated without reorientation, which may have introduced error in the sectional view density measurements of some subjects.
Expansion forces are transmitted to the midface and craniofacial complex through the zygomaticomaxillary, pterygomaxillary, and nasomaxillary resistance areas. A reactive force, called the buttressing effect, develops against the expansion forces according to patient age and the circummaxillary bone rigidity. Investigators vary in their opinions of this issue. Haas19 reports that pressure is generated by the alveolar processes, palatal vault, maxilla articulations (frontomaxillary, nasomaxillary, zygomaticomaxillary sutures), and zygomaticotemporal regions. Lines20 attributed the pressure to frontomaxillary, zygomaticotemporal, zygomaticofrontal, and zygomaticomaxillary sutures; Timms1 suspected the midpalatal suture and Revelo and Fishman21 suspected zygomatic area and maxillary articulations, while Shetty et al.22 and Lanigan and Mintz23 named the midpalatal and pterygomaxillary sutures as the major buttresses.
During our study, other sutures implicated as important contributors to RME could not be visualized as precisely as the midpalatal sutures as a result of complex regional anatomy or because the suture was too thin to trace. Therefore, we measured the bone density at several locations important in RME. The region designated as the posterior maxillary buttress includes the pterygomaxillary region. The medial maxillary buttress includes the zygomatic processes, and the anterior maxillary buttress includes the nasal area.
We were unable to find correlations between density and skeletal measurements or between density and AW or CW. Among the dental measurements, ImA was statistically highly significantly correlated (P < .01). A change in ImA may reflect alveolar bending, while tooth movement is negligible because of hyalinization within alveolar bone. Our results show a 32.1–43.3% correlation between density measurements and ImA increase. According to the calculated equation models, the ImA increase corresponding to specific skeletal density can be estimated. However, the variables were not distributed linearly, which may reflect the CT slice thickness and other factors (such as suture properties, appliance type, generated forces, and interaction between the forces and skeletal and soft tissues) that affect skeletal and dental RME outcomes. Furthermore, the number of subjects is an important limitation of this study affecting the data distribution used to formulate the model equation. Unfortunately, CT is not a routine diagnostic tool and has considerable adverse effects. The data used in this study comprised archived material obtained approximately 10 years ago when CBCT was uncommon; since then, new CT procedures have not been performed.
Bone and suture mineral densities may provide clues about the growth and resistance of skeletal components, but they are inadequate for providing measurable diagnostic information characterizing the maxillofacial area and for guiding the decision to elect SARME.
CT is a reliable method with which to measure mineral density. Bone and suture density is one component of maxillary resistance to RME and affects the ImA increase.
There is a highly significant correlation between the suture and skeletal density (HU) and the ImA increase.
However, density measurements are not the sole or definitive indicator predicting RME outcome.