Correlation Between the Extension of the Maxillary Sinus Floor and Malocclusion in Adolescent Patients Using a Cone-beam Computed Tomography

Background: The correlation between extension of the maxillary sinus oor and vertical facial skeletal patterns is important for designing orthodontic treatment plans. We correlated the extension of the maxillary sinus (MS) oor with different facial skeletal malocclusion in adolescent patients aged 10-19 years old using a cone-beam computed tomography. Methods: The relationship between the root tips of upper posterior teeth and the sinus oor was typed and scored (0-4) using cone-beam computed tomography (CBCT) images. Scores of each tooth and each patient were formulated. The cephalometric radiographs were analyzed with Dolphin software to diagnose patients' facial skeletal types into three groups according to Frankfort mandibular plane angle (FMA) as high-angle, average-angle and low-angle groups and ANB angle as Class I, Class II and Class III groups separately. The inuences caused by age, sex, and facial skeletal malocclusion on the tooth and patient score were analyzed. Results: In high-angle group, the rst molar (FM) score was signicantly higher than that in the average-angle and low-angle groups (P = 0.018). Age has a positive effect on the patient score and some of the tooth scores, such as the second premolar (SPM) score, FM score, and the second molar (SM) score, and in the skeletal facial malocclusion. With increasing age, SM score increased along with FMA angles. No signicant difference was found in the distribution of patient and tooth scores by ANB angles. Conclusion: The relationship between the maxillary posterior roots and the sinus oor is correlated with facial skeletal malocclusion in 10-to 19-year-old adolescents. More rst molar root tips were inside the sinus in the high-angle group compared with the average-angle and low-angle skeletal patterns. With advancing age, more second molar root tips were more close to contact the sinus oor in the high-angle group. Consideration of the root-sinus position of the posterior maxillary area in different skeletal malocclusion before establishing an orthodontic treatment plan is recommended. that age (10–19 years) was signicantly and positively related to the root-sinus position. Maxillary molar roots were more close the sinus oor with the increasing of age, however, no signicant sex difference was found. We also demonstrated that there was a positive relationship between roots projecting into the sinus and the high-angle skeletal malocclusion. This nding was similar to the results of Costea 21 et al., who showed that the roots of the second molars were located farther from the sinus oor in a hypodivergent biotype compared with the normodivergent and hyperdivergent facial patterns. A comparison of our ndings with previous studies is dicult since most previous studies did not consider root-sinus types within skeletal malocclusion in adolescents. Facial skeletal biotype is related to the growth of the mandible. Researches 22 have shown relationships between facial skeletal characteristics and multiple factors like occlusal forces, morphology of masticatory muscles and shape of mandible. Due to short mandibular ramus in high-angle patients, the maxillary plane will rotate and a compensatory dentoalveolar mechanism will be induced 23 three-dimensional; ANB angle: the angle between the NA (Nasion-A point) line and NB (Nasion-A point) line; CBCT: cone-beam computed tomography; FM: rst molar; FMA: Frankfort mandibular plane angle; FPM: rst premolar; MS: maxillary sinus; SM: second molar; SPM: second premolar; WHO: World Health Organization.


Background
The maxillary sinus (MS) is the largest paranasal sinus and the rst to develop, and it grows proportionally with facial bones 1,2 . The inferior wall of MS, or sinus oor, supports the roots of the maxillary dentition and continues with the alveolar process. The relationship between the root apex and the MS oor (root-sinus relationship), and the correlation between the root-sinus position and facial skeletal malocclusion are less considered when planning orthodontic treatment. However, tooth movement through the MS, is regarded as one of the most di cult tasks in orthodontics 3 . Except the side effects like root resorption and pulp vitality, movement of the tooth against the cortical bone is another challenging problem to address. Hence, the root-sinus relationship is relevant in orthodontic clinic.
The World Health Organization (WHO) de nes adolescents to those between 10 and 19 years of age 4 . During adolescence, the growth and development of the MS are associated with the posterior maxillary teeth, which can be susceptible to malocclusion 5,6 . There is a close relationship between the upper maxillary teeth and the sinus oor [7][8][9] . However, few studies have addressed the correlation between the extension of the MS oor and skeletal malocclusion in adolescents.
Assessing the anatomical position of posterior maxillary tooth roots to the MS oor is essential when they are in close proximity. Conebeam computed tomographic (CBCT) imaging is indicated to provide detailed information at the structure of interest with relatively low radiation exposure 10 , which can also be used to analyze three-dimensional (3D) cephalometric images and create tissue reconstruction for further analysis. Our goal was to correlate the extension of the MS oor to malocclusion in adolescents and to provide recommendations for clinical orthodontic treatment.

Subjects
The cephalometric radiographs and dental CBCT images used in this study were selected from the CBCT database at radiology department at the Stomatological Hospital of ***. This study was approved by an institutional review board for its retrospective property. Patients with pre-obtained CBCT images for orthodontic needs within the recent 2 years were included and a total image data of 178 patients were nally analyzed in this study ( Figure 1).

Radiographic Evaluations
Cephalometric data were saved in JPG format, and CBCT data were saved in DICOM format. Evaluations of facial skeletal malocclusion and root-sinus relationship were processed using 3D software (Ver. 11.7 Premium; Dolphin Imaging, Chatsworth, CA, USA).
The following anatomic landmarks were identi ed on lateral cephalometric radiographs to determine the classi cation of skeletal malocclusion: Sella, Nasion, Porion, Orbitale, point A, point B, Pognion, Menton, and Gonion. ANB is de ned as the angle between the NA (Nasion-A point) line and NB (Nasion-A point) line. Based on the range of the ANB angle 11,12 , patients were divided into three malocclusion groups: 0°< ANB <4° were classi ed into skeletal Class I malocclusion, ANB ≥ 4° were classi ed as skeletal Class II malocclusion, and ANB < 0° were classi ed as skeletal Class III malocclusion. The line connecting Menton and Gonion was identi ed as the mandibular plane, and the angle formed by the mandibular plane and Frankfort horizontal plane was the Frankfort mandibular plane angle (FMA). A normal value was considered to be 25° ± 3°1 3 . Patients were divided into average-angle (22° ≤ FMA ≤ 28°), low-angle (FMA < 22°), and high-angle (FMA > 28°) skeletal malocclusion types.
The root tip was precisely located with CBCT images in coronal, sagittal, and transverse sections and the root-sinus relationship was veri ed according to Jung 14 . The vertical spatiality between the MS oor and molar roots could be classi ed into four types: 0) no contact between the root and the cortical borders of the MS; 1) slight contact between the root and the cortical borders of the MS; 2) a lateral protrusion of a root in the sinus cavity without the projection of its apex into the sinus; and 3) a projection of the root apex into the sinus cavity ( Figure 2). The relationship between the root tips of the premolars as well as molars and the sinus oor was analyzed on CBCT scans and repeated after 30 days to assess the reliability. The left and right sides of each patient, each root tip of each premolar or molar, were categorized and regarded as independent. The examiner was blinded to the basic information of patients and all previous measurements.
Another examiner conducted all cephalometric measurements and repeated them after 30 days. The examiner was also blinded to the details of patients and previous evaluation results.

Statistical analysis
For each tooth, an average tooth score was calculated according to Formula 1 and recorded as FPM (the rst premolar), SPM (the second premolar), FM (the rst molar), and SM (the second molar) separately. For each patient, an average patient score was calculated using Formula 2.
Results are presented in numbers (percentages), means, and standard deviations. All statistics were performed with IBM SPSS Statistics 23.0 software. A normality test was used to check the distribution of groups, and a consistency test was used to determine the test-retest reliability. For variables that were not normally distributed, differences were analyzed by the Kruskal-Wallis test. These included age, teeth per patient, roots per patient, tooth score, and patient score. Chi-square tests were used to analyze the differences between groups according to sex. Correlations between patient score, tooth score, age, and skeletal malocclusions were analyzed with the Spearman coe cient of correlation. The in uence of skeletal malocclusion on tooth and patient scores was analyzed using a general linear regression model. P<0.05 was considered to be statistical signi cance.

Results
This research included 178 patients, consisting of 94 female patients and 84 male patients. The consistency test results for the reliability of cephalometric analysis and sinus-root type assessment between test-retest were ĸ = 0.87 (P < 0.001) and ĸ = 0.79 (P < 0.001), respectively.
The distributions of gender, age, teeth, and roots per patient, according to FMA angles and ANB angles, are given in Table 1. A comparison of gender, age, roots, and teeth per patient by FMA angles showed no signi cant differences between different sagittal skeletal pattern groups (P > 0.05). For ANB angles, a signi cant difference in teeth per patient was found between different coronal skeletal pattern groups (P < 0.05). The types of root-sinus relationship in the upper posterior area by ANB angles and FMA angles are presented in Table 2. A total of 2569 upper maxillary posterior roots (675 premolars and 582 molars) were assessed. These consisted of 791 type 0 roots, 421 type 1 root, 453 type 2 roots, and 904 type 3 roots. The most frequent root type was type 3 (35.19%), and the second most frequent was type 0 (30.79%).
Grouped by ANB angles, 996 roots were categorized into Class I malocclusion, 1283 roots into Class II malocclusion, and 290 roots into Class III malocclusion. Grouped by FMA angles, 1108 roots were categorized into average-angle malocclusion, 566 roots into low-angle malocclusion, and 895 roots into high-angle malocclusion.  (100) The distribution of root-sinus types of each posterior teeth roots in different skeletal malocclusion groups is shown in Table 3. The most frequent root-sinus type in all the roots of all the rst premolars was type 0 regardless of malocclusion groups. Type 3 was the dominant relationship in the majority of roots of upper molars. No signi cant difference was found in the distribution of patient score and tooth score by ANB angles. Signi cant differences were found in the FPM and FM score by FMA angles ( Table 4). The FPM in high-angle group was signi cantly lower than that in the average-angle and low-angle groups (P = 0.039). And FM in the high-angle group was signi cantly higher compared with that in the average-angle and low-angle groups (P = 0.018).  When analyzed with a general linear regression model (Table 5a and Table 5b), age was taken as continuous variables, and patient score, FPM score, SPM score, FM score, SM score as dependent variables. After controlling for age, signi cant differences were found between the ANB angle groups as well as the FMA angle groups for patient score, SPM score, FM score, and SM score. Age was associated with patient score and in some of the tooth scores, such as SPM score, FM score, and SM score, when considering the skeletal facial pattern. tail), and SM score (Spearman's rho [ρ] = 0.195; P = 0.001, two tail) were signi cantly and positively correlated with age. Interestingly, FM score was not related to age without considering facial skeletal discrepancy. SM score was correlated with FMA angles (Spearman's rho [ρ] = 0.155; P = 0.041, two-tail test), indicating that more root apexes of the upper second molars were likely to penetrate into the sinus in the high-angle group. There was no signi cant correlation between sex and either tooth score or patient score (data not shown).

Discussion
There is a correlation between the root tips of upper posterior teeth and the extension of the MS oor in the skeletal malocclusion groups in adolescents. FMA angles seemed to be related to the root-sinus position. The proximity of the upper second molars roots to the sinus oor was positively related to the FMA angles, indicating that vertical facial skeletal malocclusion might in uence the root-sinus relationship in the posterior maxillary area. We found that ANB angles were less related to root-sinus relationships, which suggests that in adolescents, the proximity of the posterior maxillary teeth to the sinus oor might be independent of the anteroposterior position of maxilla and mandible.
The root-sinus relationship has been studied in different age groups 7,8,14−19 . Ok et al 7 reported that the relationship between the sinus oor and the posterior maxillary teeth differed with age. The results showed that the roots of the maxillary rst and second molars drifted apart from the sinus oor with the increasing of age, especially in the fourth decade. Tian 19 et al. con rmed that the proximity between the posterior maxillary teeth and the MS varied with age. And age less than 40 years showed a higher likelihood of roots above or inside the sinus oor. Park 20 noted a signi cantly closer distance between maxillary root tips and the sinus oor with a high-angle skeletal pattern and a sizeable gonial angle in older (20-28 years) male groups than younger (10-20 years) female groups. These results were partly consistent with our study. We found that age (10-19 years) was signi cantly and positively related to the root-sinus position. Maxillary molar roots were more close the sinus oor with the increasing of age, however, no signi cant sex difference was found. We also demonstrated that there was a positive relationship between roots projecting into the sinus and the high-angle skeletal malocclusion. This nding was similar to the results of Costea 21 et al., who showed that the roots of the second molars were located farther from the sinus oor in a hypodivergent biotype compared with the normodivergent and hyperdivergent facial patterns. A comparison of our ndings with previous studies is di cult since most previous studies did not consider root-sinus types within skeletal malocclusion in adolescents.
Facial skeletal biotype is related to the growth of the mandible. Researches 22 have shown relationships between facial skeletal characteristics and multiple factors like occlusal forces, morphology of masticatory muscles and shape of mandible. Due to short mandibular ramus in high-angle patients, the maxillary plane will rotate and a compensatory dentoalveolar mechanism will be induced 23 . Moreover, there seemed to be thinner cortical bone plates in hyperdivergent patients compared with normodivergent and hyperdivergent patients 24,25 . Thus, the sinus will expand because of weak cortical bone block, and roots of maxillary posterior teeth are more likely project into the sinus because of the absence of posterior dentoalveolar compensation. This may be to some extent explain why more molar roots would contact or inside the sinus in patients with high-angle skeletal malocclusion.
Factors including facial pattern, teeth crowding, and molar relationship, are involved in orthodontic treatment plans. The relationship between the MS oor extension and malocclusion also affects the orthodontic plan such as the intrusion of the maxillary posterior teeth.
We showed that more than one-third of maxillary root tips in both Class II and the high-angle skeletal group ( precautions for clinicians at the beginning of dealing with adolescent orthodontic patients. As a 2-dimensional (2D) radiographic imaging technique, panoramic radiography can be used to roughly assess the relationship between the MS oor and the posterior root apex. However, the reliability is uncertain due to the superposition and magni cation of the anatomical structure 40 . Compared with traditional panoramic radiographs, CBCT images can provide additional information and reduce error, making it useful for orthodontic diagnosis. For speci c clinical scenarios, such as unerupted teeth, root resorption, and airway problems, the additional information provided by CBCT can result in a more clinically orientated treatment compared to conventional imaging [41][42][43][44][45] . In our study, 2D cephalometric radiographs were used to measure the ANB angle and the FMA angle. CBCT was used to assess the relationship between the posterior maxillary roots and the sinus oor at different orientations using Dolphin software. CBCT imaging provides unique features and advantages over conventional extra-oral radiographic imaging such as 3D visualization and virtual 3D model reconstruction 46 . In addition to these advantages, CBCT data can also be used for 3D cephalometric analysis. Obtaining ANB angles and FMA angles through 3D cephalometry is more precise and e cient. However, the large eld of view (FOV), which covers most of the craniofacial skeleton, is recommended for 3D cephalometric analysis. Notably, when using CBCT imaging in orthodontics, examinations should not be routinely applied but should only be used on a case-by-case basis 47 after considering the speci c clinical situation and the needs of the patient.

Conclusion
The relationship between the maxillary posterior roots and the sinus oor is correlated with facial skeletal malocclusion in 10-to 19-yearold adolescents. We identi ed more rst molar root tips inside the sinus in the high-angle group compared to the average-angle and lowangle groups. In the high-angle malocclusion, more second molar root tips contacted the sinus oor as the patient's age increased.
Consideration of the root-sinus position of the posterior maxillary area in different skeletal malocclusion before establishing an orthodontic treatment plan is recommended.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed in this study are available from the corresponding author on reasonable request.

Competing interests
There is no nancial or non-nancial competing interests.

Funding
This study was supported by Natural Science Foundation of Anhui Province (No.1908085MH255). The funding body fully refrained from exerting in uence on any aspect of the study, neither regarding the design nor the data collection nor processing. Figure 1 Flow chart of excluded and included patients Figure 2 CBCT images show the vertical spatiality between the maxillary sinus oor and molar root apex: A.0, no contact between the root and the cortical borders of the maxillary sinus; B.1, slight contact between the root and the cortical borders of the maxillary sinus; C.2, the lateral protrusion of a root in the sinus cavity without the projection of its apex into the sinus; D.3, the projection of the root apex into the sinus cavity.