Evaluation of bone area in the posterior region for mandibular molar distalization in class I and class III patients

The aim of this study was to investigate the bone area in the posterior region, which is important for mandibular molar distalization in skeletal class I and class III individuals with normodivergent and hyperdivergent vertical growth patterns. In this retrospective study, cone-beam computed tomography (CBCT) scans of 120 individuals divided into 4 groups as class I normodivergent (group-I), class I hyperdivergent (group-II), class III normodivergent (group-III), and class III hyperdivergent (group-IV). Retromolar area at crown level measurements was performed on CBCT-derived panoramic radiographs and axial sections. Retromolar area at the root level was measured 2 mm, 4 mm, 6 mm, 8 mm, and 10 mm apical to cemento-enamel junction at CBCT axial sections. Retromolar area decreased towards the root apex in all groups and smallest retromolar area was level of CEJ10mm in all groups. At any root level, the distal root of the mandibular second molar tooth was in contact with the mandibular inner or outer lingual cortex; 50% in group-I, 46.7% in group-II, 23% in group-III, and 23% in group-IV. In normodivergent individuals, the retromolar area length at the root level is observed to be higher in class III than in class I at almost every level. In hyperdivergent individuals, on the other hand, only at CEJ10mm level, it is higher in length in class III than in class I. Vertical growth pattern has no effect on the root and crown level retromolar area in class I and class III individuals. CBCT provides more useful information than panoramic radiographs for patients who are scheduled for large mandibular molar distalization.

In recent years, with the usage of miniscrews and miniplates, camouflage treatment with distalization of the mandibular arch can be applied to patients who do not want orthognathic surgery and have moderate skeletal incompatibility [8][9][10][11].
Before distalization of the mandibular molar, the hard and soft tissues in the region should be evaluated. For distalization, sufficient retromolar area and adequate attached gingiva are required distal to the mandibular second molar tooth [12]. Studies on two-dimensional radiographs reported that the border of the retromolar area is the anterior edge of the ramus. However, studies on cone-beam computed tomography (CBCT) images in recent years showed that this border is the lingual cortex of the mandibular corpus [13]. Consequently, retromolar area values in crown level determined in panoramic images may misguide the clinician. The preferred reference retromolar area values before the treatment need to be taken from root level adjacent the lingual cortex of the mandibular corpus. There is a lack of studies in the literature regarding the effect of both sagittal and vertical facial growth patterns on the retromolar area.
The aim of this study was to evaluate the bone area in the posterior region, which is important in mandibular molar distalization, in skeletal class I and class III individuals with normodivergent and hyperdivergent vertical growth patterns. The null hypothesis was that the bone area in posterior region for mandibular molar distalization would not be different among patients with different sagittal and vertical skeletal patterns.

Materials and methods
This retrospective study was carried out with the approval (Protocol no: 2019/274) from the Inonu University Health Sciences Non-Interventional Clinical Research Ethics Committee. Among the 2500 individuals whose CBCT images were scanned, 850 records were noted as meeting the inclusion criteria. These 850 images were then reevaluated based on the exclusion criteria. After the exclusion of images that did not meet the criteria, the remaining 120 records were divided into four groups of 30 individuals in each group: class I normodivergent (group-I), class I hyperdivergent (group-II), class III normodivergent (group-III), and class III hyperdivergent (group-IV).
Studies have shown that the GoGn-SN measurement has the highest accuracy in classifying vertical facial types. Additionally, the FMA measurement has the highest correlation with the GoGN-SN measurement [14]. Inclusion criteria: skeletal class I malocclusion (0°≤ANB≤4°), skeletal class III malocclusion (ANB<0°), normodivergent facial profile (32°≤GoGn-SN≤38° or 25°≤FMA≤29°), hyperdivergent facial profile (GoGn-SN>38° or FMA>29°), individuals between the ages of 18-30. Exclusion criteria: missing teeth other than third molars, erupted third molars, pathology of the third molar germ, crowding of the mandible more than 5 mm, history of orthodontic treatment, facial asymmetry, unhealthy periodontal state with noticeable alveolar bone loss, craniofacial trauma or syndrome, mandibular posterior region pathology, abnormal dental conditions at mandibular molars (dilacerations, idiopatic root resorptions, c-shaped root canals, severe rotations, severe enamel losses etc.), artifacts in images, and patients without centric occlusion. Radiographic images were obtained using a flat panel CBCT device from Newtom 5G (Newtom QR DVT 9000 Quantitative Radiology, Verona Italy). The device works with 1-20 mA and 110 kVp value as standard, with cone beam technique. The voxel size is 0.25 mm and the scan time is 18 s. CBCT data were transferred to the NNT (New New Tom) software program and the slice thickness was adjusted to 0.25 mm. For image reorientation, the coronal plane passes anteriorly from the midpoint of the incisal edge of the two mandibular central incisors, parallel to the transporionic line (connecting the upper central points of right and left external auditory meatuses), and the midsagittal plane passes through the middle of the Anterior Nasal Spina, Crista Galli, and two mandibular central incisors. The mandibular occlusal plane passes through the midpoint of the incisal edge of the two mandibular central incisors, posteriorly from the mesiobuccal cusp of the right and left mandibular first molars [13].
Retromolar area at crown level (PAN 47-ABR ) was measured on CBCT-derived panoramic radiographs. This area is the distance between the anterior edge of the ramus and the distal right mandibular second molar, parallel to the mandibular occlusal plane at the occlusal level ( Fig. 1).
All measurements on CBCT images were performed at the root and crown levels on axial sections taken parallel to the mandibular occlusal plane. In axial sections, the retromolar area at crown level (AX 47-ABR ) is the shortest distance between the distal mandibular right second molar and the anterior border of the ramus. This area was measured on the axial sections corresponding to the marked point on the panoramic images and parallel to the posterior occlusal line. Posterior occlusal line: it is the line passing through the buccal cusp crests of the right mandibular first and second molars (Fig. 2).
In the measurements at root level in axial sections, the shortest distance between the most lingual point of the distal root of the right mandibular second molar tooth and the inner and outer lingual cortex of the mandibular corpus was measured to be parallel to the posterior occlusal line. From the cemento-enamel junction to the apical, 10 measurements were made at 5 levels: CEJ 2mm , CEJ 4mm , CEJ 6mm , CEJ 8mm , and CEJ 10mm .
Distal roots of some molars were found to be in contact with the inner or outer lingual cortex of the mandible (Fig. 2). Values at these root levels in contact were accepted as "0 mm." All of the images, another measurement was performed by another researcher 2 weeks after the first measurement. Intra-class correlation coefficients analysis was used to evaluate method error (0.90-0.99).

Statistical analysis
In the power analysis, it was calculated that 30 individuals in each group should be included in the study at 80% power with an effect size of 0.31, 5% error tolerance, and 95% confidence interval. The data obtained as a result of the study were analyzed with IBM SPSS V23 software. Conformity to normal distribution was evaluated with the Kolmogorov Smirnov test. Oneway analysis of variance test was used for comparisons between groups. Examinations were made with Tukey HSD test, one of the multiple comparison tests. Intragroup changes were analyzed by repeated measures analysis of variance. The analysis of categorical data was performed with the chi-square test. The significance level was evaluated as P<0.05 in all analyses.

Results
In the study, there were 15 females and 15 males in group-I, 16 females and 14 males in group-II, 13 females and 17 males in group-III, and 15 females and 15 males in group-IV. The mean ages of the individuals in the groups were 21.7, 20.4, 19.4, and 18.5 years, respectively.
PAN 47-ABR values were found to be approximately 3 mm less than AX 47-ABR values in all groups. The retromolar area at the crown level in axial images was significantly higher in length group-III and group-IV than group-I and group-II (P<0.001). Group-III individuals were found to have significantly more retromolar area in other measurements than group-I individuals, except for CEJ 2mm-inner , CEJ 2mm-outer , CEJ 6mm-outer , and CEJ 8mm-inner . Group-IV individuals had significantly more retromolar area than group-II individuals only in CEJ 10mm-inner (P=0.011) and CEJ 10mm-outer (P=0.001) measurements ( Table 1).
The distances from the inner and outer lingual cortex at root levels decreased from coronal to apical in all groups. At the CEJ 10mm level, the distance between the distal root of the mandibular second molar tooth and the inner lingual cortex of the mandibular corpus was 1 mm in group-I, 1.1 mm in group-II, 2.2 mm in group-III, and 2.4 mm in group-IV. At the CEJ 10mm level, the distance between the inner lingual cortex and the outer lingual cortex of the mandibular corpus was 3 mm in group-I, 2.4 mm in group-II, 4.3 mm in group-III, and 3.6 mm in group-IV (Table 2). It was found that the contact of the distal root of the mandibular second molar tooth with the inner lingual cortex of the mandible was seen at the level of the first CEJ 4mm and increased towards the apex. In group-I, group-II, and group-III, roots in contact with the outer lingual cortex of the mandible are only at the level of the CEJ 10mm . Roots in contact with the outer lingual cortex of the mandible were observed at the first CEJ 6mm level in group-IV (Table 3).
In 50% of group-I, 46.7% of group-II, 23% of group-III, and 23% of group-IV, contact status was detected between the distal root of the mandibular second molar tooth and the inner or outer mandibular lingual cortex at any level (Table 4).

Discussion
The distalization of mandibular molar teeth is one of the difficult treatments in orthodontics [4]. The factors, such as the dense cortical bone structure, the limitations in the placement of temporary anchor devices, and the anatomical structures that may limit tooth movement, complicate the distalization of mandibular molars [8,13,15].
Dang et al. [9] stated that in CBCT, the roots of the lower second molars are in contact with the inner lingual cortex of the mandibular corpus, and the limits of the distalization movement should be investigated in order to prevent this. Lateral cephalograms provide limited information regarding Table 1 Comparison of retromolar area values at crown levels (PAN 47-ABR , AX 47-ABR ) and root levels (CEJ 2mm,4mm,6mm,8mm,10mm ) between groups Sd, standart deviation; mm, milimeter; ABR, anterior border of ramus; PAN 47-ABR , retromolar area at crown level on panoramic images; AX 47-ABR , retromolar area at crown level on axial CBCT sections; CEJ, cemento-enamel junction. *One-way variance analysis. a-c, there is no statistical difference between groups with the same letter  the retromolar area and root contact with the inner lingual cortex [13]. Kim et al. [13] mentioned that the limit of mandibular molar distalization is not the anterior of the ramus, but the lingual cortex of the mandibular corpus, and emphasized the importance of 3D imaging. In our study, the anatomical limits of molar distalization in the mandibular jaw in individuals with different sagittal and vertical growth patterns were determined. For this purpose, we focused on the bone structures in the retromolar area that may limit the distal movement of the right mandibular second molar tooth. Only hyperdivergent and normodivergent individuals were included in our study because sufficient number of class III hypodivergent individuals could not be reached in approximately 2500 scanned CBCT images. This may be due to the fact that many individuals with skeletal class III malocclusion have a vertical growth pattern. Mandibular molar distalization is an orthodontic camouflage treatment method that is generally preferred in skeletal class I or class III patients with angle class III dental relationship. Therefore, as in similar studies, class II individuals were not included in our study [13,16].
Most of the studies evaluating the effects of different vertical and sagittal growth patterns on the retromolar area were performed on 2D radiographs and to observe the eruption path of impacted teeth [17][18][19][20]. Choi et al. [16] reported that class III normodivergent individuals had significantly more retromolar area than class I normodivergent individuals in their retromolar area measurements only at the furcation level on CBCT axial sections. In our study, it was observed that the retromolar area in normodivergent class III individuals at most root levels was significantly larger than in normodivergent class I individuals. Soft tissue distal to the mandibular second molar can restrict the amount of distalization movement. After the distalization movement, the distobuccal surface of the mandibular second molar tooth should be located within the borders of the attached gingiva [12]. Since our study is a retrospective study on CBCT sections, it is impossible to look at soft tissue thickness. So, this situation was not taken account.
Mandible morphology of hyperdivergent class I and class III individuals may be different. In our study, the retromolar area in hyperdivergent class III individuals was higher in length at most levels than hyperdivergent class I individuals, but this excess was found to be significantly higher only at the CEJ 10mm level.
Kim et al. [13] emphasized that different vertical growth patterns may affect the remodeling of the mandibular ramus and thus the retromolar area, and this issue should be taken into account. Zhao et al. [21] reported that class I hyperdivergent patients had the smallest retromolar area in their study, which they performed on CBCT sections in only class I patients. In our study, it was seen that the retromolar area Table 3 Distribution of root numbers in contact with inner or outer mandibular lingual cortex according to root levels (CEJ 2,4,6,8,10 mm ) n, number of roots in contact with the inner or outer mandibular lingual cortex; mm, milimeter; %, the ratio of the teeth whose root is in contact with the mandibular lingual cortex at that level to the total number of teeth measured at that level; CEJ, cemento-enamel junction. *Chi-square test

Root levels
Group I n (%) Group II n (%) Group III n (%) Group IV n (%) p* Table 4 General contact ratios between root and mandibular lingual cortex by groups n, number of teeth in the group whose root is in contact with the inner or outer mandibular lingual cortex at any level. %, the ratio of the teeth in contact with the inner or outer mandibular lingual cortex to the total number of teeth in that group. *Chi-square test in normodivergent individuals was higher in length than in hyperdivergent individuals in both class I and class III individuals at the root level, but this excess did not have a significant impact on the retromolar area. This finding shows that vertical growth pattern does not have a significant impact on the retromolar area. Future studies addressing the 3D mandible morphology of patients with different vertical growth patterns will make significant contributions to this issue. In this study, intra-group measurements decreased towards the root apex and the lowest values were obtained at the CEJ 10mm level in all groups. Kim et al. [13] also stated that in class I normodivergent individuals, similar to our study, the retromolar area at the root level decreases towards the apex. Choi et al. [16], on the other hand, stated that the retromolar area decreased towards the root apex only in class III individuals, but there was no such decrease in class I. The retromolar area towards the root apex may be reduced due to differences in inclination between the lingual cortex of the mandibular corpus and the mandibular second molar tooth towards the apex in the buccolingual direction. Specifically, the lingual cortex may be more inclined towards the apex compared to the mandibular second molar tooth.
Although there are studies on the incline of the mandibular corpus [22,23] in the literature, there are no studies specifically examining the incline of the lingual cortex.
When the tooth root comes into contact with the inner lingual cortex of the mandible, tooth movement slows down and the possibility of root resorption increases [13,[24][25][26]. If the movement is continued after contact with the inner lingual cortex, contact with the outer lingual cortex will occur, increasing the possibility of alveolar bone resorption in the region. In severe cases, perforation of the cortex, exposure of the root surface, and gingival recession may occur. As a result, the periodontal support of the region is lost [13]. Based on all this information, the limits of molar distalization at the CEJ 10mm level in the mandible can be determined.
Choi et al. [16] found that the distance of the distal root of the second molar tooth from the inner lingual cortex in their measurements from the 6 mm apical furcation was found to be greater in class I than class III ones, which is not in line with the findings of our study. Kim et al. [13] determined the distance between the root and the inner and outer lingual cortices to be 2.87 mm and 6.73 mm, close to the root apex, in class I normodivergent individuals, and they stated that the amount of molar distalization to be planned as approximately 3 mm would be reasonable. In our study, these values were determined as 1 mm and 4 mm in class I normodivergent individuals, respectively. Although Kim et al. [13] stated that the existence of mandibular third molars does not have a significant effect on the retromolar area, there are many studies [18,19,27,28] indicating that the retromolar area is more in individuals whose third molars have erupted into the mouth. Therefore, individuals with erupted third molars were not included in our study in order to ensure standardization and to eliminate the possible effects of these teeth. Impacted third molars were not considered exclusion criteria as they did not affect our measurements. However, it is recommended that these teeth be extracted before the distalization phase from a clinical point of view. The inclusion of individuals with erupted third molars in their study can be considered the reason why the retromolar area length values in the study of Kim et al. [13] were higher than the values in our study.
It is recommended that in patients with a need for mandibular molar distalization greater than 3 mm, it would be more appropriate to obtain space by premolar tooth extraction [12]. In some individuals, the roots of the molar teeth may be in contact with the inner or outer lingual cortex before starting the distalization movement. In the axial sections we examined, especially in class I individuals, a high rate of root in contact with the inner lingual cortical bone was detected. It was seen that the highest contact rate in all groups was at the CEJ 10mm level. Similar to our study, Kim et al. [13] found the highest contact rate at the root apex level. This can be explained by the reduction of the retromolar area towards the root apex. While the rate of teeth in contact with the inner lingual cortex was high, the outer lingual cortex was found to be very low in all groups. Kim et al. [13] found the incidence of contact between the inner-outer lingual cortex and the root to be 35.3% in class I normodivergent individuals, and it is thought that the reason why it was seen as 50% in our study may be due to interracial mandibular morphology differences.
It is predicted that the distal root of the mandibular second molar will make contact with the inner cortical bone after an average of 1 mm distalization movement in skeletal class I individuals and 2.2 mm in skeletal class III individuals. After this stage, more controlled forces should be given, since the probability of resorption at the root will increase. The mandibular molar distalization limits are approximately 3.4 mm in skeletal class I individuals and approximately 6 mm in skeletal class III individuals. If more than these amounts of distalization movement are made, external cortical bone perforation may occur and the tooth root may be exposed. In cases where more distalization movements are planned than these amounts, it is recommended to take CBCT before treatment.

Conclusions
The main findings were the following: • The amount of distalization of the mandibular molar tooth is affected by the retromolar area at the level of the