CBCT-Based Three-Dimensional Phenotyping of Skeletal Class II Malocclusion in Yemeni Adults: A Multivariate Workflow for AI-Ready Orthodontic Diagnostics

CBCT-Based Three-Dimensional Phenotyping of Skeletal Class II Malocclusion in Yemeni Adults: A Multivariate Workflow for AI-Ready Orthodontic Diagnostics | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article CBCT-Based Three-Dimensional Phenotyping of Skeletal Class II Malocclusion in Yemeni Adults: A Multivariate Workflow for AI-Ready Orthodontic Diagnostics Salah M. Ben Hafedh¹, Ghamdan Al‑Harazi², Ramy Ishaq³ This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7880628/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Skeletal Class II malocclusion is among the most common orthodontic problems, showing substantial variation in its craniofacial presentation. Objective: To identify distinct three-dimensional craniofacial phenotypes of Class II malocclusion in a Yemeni adult population using CBCT and multivariate analysis. Methods: CBCT scans from Yemeni adults were analyzed. Linear and angular cephalometric parameters were extracted. Principal component analysis (PCA) and cluster analysis (CA) were applied to derive phenotypes. Results: Five distinct phenotypes were identified, reflecting variability in sagittal, vertical, and transverse skeletal parameters. Canonical discriminant analysis confirmed robust separation. Conclusion: CBCT-based phenotyping highlights the heterogeneity of Class II malocclusion. Findings can guide individualized treatment planning and provide a framework for integrating AI and genetics into orthodontic diagnostics. Class II malocclusion CBCT cephalometrics PCA cluster analysis phenotyping AI genetics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1. Introduction Skeletal Class II malocclusion is one of the most common orthodontic problems worldwide, accounting for up to one-third of orthodontic patients, and is characterized by a wide spectrum of craniofacial discrepancies involving sagittal, vertical, and transverse dimensions [ 1 , 2 ]. Traditional diagnostic approaches, such as the Angle classification and the Wits appraisal, have been widely used to define this condition, but they provide limited insight into its heterogeneous skeletal basis [ 2 , 3 ]. Cephalometric methods, beginning with the pioneering work of Broadbent and Steiner, have allowed orthodontists to quantify skeletal and dental relationships more systematically [ 5 , 6 ]. However, two-dimensional cephalograms still suffer from projection errors, landmark superimposition, and limited representation of craniofacial complexity [ 7 , 8 ]. The introduction of cone-beam computed tomography (CBCT) has transformed craniofacial imaging by providing isotropic three-dimensional data, allowing for accurate assessment of skeletal and dental structures with relatively low radiation compared to medical CT [ 19 , 29 ]. CBCT-based morphometric analyses have revealed population-specific variations in craniofacial form, including differences between Western, East Asian, and Middle Eastern groups [ 23 , 24 , 31 ]. Such population-based studies highlight the importance of establishing localized cephalometric standards before extrapolating diagnostic criteria to diverse ethnic cohorts [ 24 , 33 ]. For the Yemeni population, normative data remain scarce, with only a few studies exploring malocclusion prevalence and craniofacial morphology [ 33 , 40 ]. Despite improved imaging, the biological complexity of Class II malocclusion remains unresolved. Early studies demonstrated that some patients present with mandibular retrusion, while others show maxillary protrusion, vertical dysplasia, or combined features [ 7 , 9 ]. Longitudinal studies in untreated Class II subjects have highlighted growth-related variability and the influence of skeletal maturation on treatment outcomes [ 9 , 35 ]. Appliance-based interventions such as the Herbst appliance further demonstrated that mandibular repositioning alone cannot explain the diversity of Class II phenotypes [ 10 ]. Therefore, a more robust multivariate approach is required to delineate clinically meaningful subgroups. Principal component analysis (PCA) and cluster analysis (CA) have been increasingly employed in orthodontics to address this challenge. These methods reduce dimensionality by identifying patterns of correlated craniofacial variables and grouping individuals into homogeneous clusters [ 18 , 26 ]. Uribe et al. [ 11 ] were among the first to apply multivariate statistics to moderate-to-severe Class II patients, reporting distinct skeletal subtypes in a North American cohort. More recent work has extended such analyses into different populations using CBCT-derived measurements, uncovering novel phenotypes that may be obscured in two-dimensional studies [ 18 , 23 ]. Canonical discriminant analysis (CDA) further validates these clusters by quantifying the accuracy of group separation [ 18 ]. The application of artificial intelligence (AI) has added new dimensions to this field. Machine learning algorithms have been successfully applied to cephalometric landmark detection, phenotype recognition, and treatment outcome prediction [ 12 , 13 , 20 ]. Studies using convolutional neural networks and ensemble models have achieved landmark localization accuracy comparable to human examiners, thereby streamlining cephalometric workflows [ 13 , 22 ]. Beyond diagnosis, AI has been used to predict treatment success for appliances, anchorage needs, and extraction decisions [ 15 , 25 , 27 ]. These developments suggest that AI-based clustering and phenotyping could soon complement or even replace traditional statistical approaches [ 21 , 32 ]. Parallel to imaging advances, genetic studies have uncovered associations between craniofacial morphology and specific gene polymorphisms. Genome-wide association studies (GWAS) have identified candidate loci linked to mandibular retrognathism and vertical growth patterns [ 16 , 34 ]. Integrating AI with genetic datasets has been proposed as a way to better model the complex interplay between environmental and hereditary factors in malocclusion [ 17 , 28 ]. Such interdisciplinary efforts are reshaping orthodontic research, with implications for personalized treatment planning. In Middle Eastern populations, where Class II malocclusion is highly prevalent, studies remain limited [ 24 , 31 ]. Yemeni data, in particular, are underrepresented in the orthodontic literature. The lack of standardized CBCT-based phenotypic profiles hinders both clinical decision-making and the ability to compare local findings with global cohorts [ 33 , 40 ]. This gap is clinically significant because treatment protocols developed in Western or Asian populations may not be directly applicable to Yemeni patients, whose craniofacial growth patterns are influenced by unique genetic and environmental factors. Therefore, this study aimed to characterize three-dimensional craniofacial phenotypes of Class II malocclusion in a Yemeni adult population using CBCT and multivariate statistical methods. By applying PCA and CA to comprehensive cephalometric datasets, we sought to identify reproducible skeletal subgroups that reflect the inherent diversity of Class II malocclusion. We also compared our findings with international studies, including those integrating AI and genetic approaches, to highlight the potential role of interdisciplinary diagnostics in orthodontics [ 11 , 12 , 16 – 18 , 21 , 23 , 25 , 32 , 34 ]. Ultimately, our goal was to establish a foundation for individualized orthodontic treatment planning and to encourage the integration of Yemeni data into global discussions on craniofacial phenotyping. 2. Methods This was a retrospective observational study conducted at the Faculty of Dentistry, Sana’a University. The research protocol was reviewed and approved by the Research Ethics Committee of the Faculty of Dentistry, Sana’a University (Approval No. 2023/32, Date: 15 January 2023). All procedures followed the ethical standards of the Declaration of Helsinki. Written informed consent had been obtained at the time of radiographic imaging for clinical purposes, and data were anonymized before analysis [ 1 , 4 ]. CBCT scans were retrieved from the departmental archives. Inclusion criteria were: Yemeni adults aged 18–30 years; skeletal Class II malocclusion confirmed by ANB angle > 4° and Wits appraisal > 2 mm; full permanent dentition (excluding third molars); and absence of previous orthodontic or orthognathic treatment. Exclusion criteria were: craniofacial syndromes, cleft lip/palate, significant facial asymmetry, or poor image quality [ 2 , 3 , 7 , 8 ]. Based on these criteria, a final sample of n = 120 patients (56 males, 64 females) was analyzed. All CBCT scans were obtained using (PaX‑Fle × 3D P2, Ver. 1.0.0, Vatech, Korea) with a standardized protocol: field of view 15 × 1 cm, 0.3 mm isotropic voxel size, 90 kV, 10 mA, and 17-second exposure. Patients were positioned with the Frankfort horizontal plane parallel to the floor and teeth in maximum intercuspation during scanning [ 19 , 29 ]. This standardized acquisition ensured reproducibility and comparability with previous CBCT-based craniofacial studies [ 23 , 24 ]. Three-dimensional cephalometric analysis was performed using Dolphin Imaging software (Dolphin Imaging & Management Solutions, Chatsworth, CA, USA). A total of 36 linear and angular measurements were extracted, covering sagittal, vertical, and transverse dimensions. Landmarks were identified according to conventional cephalometric definitions established in previous literature [ 5 , 6 , 9 , 10 ]. To reduce inter-observer variability, all landmarks were identified twice by the principal investigator and re-evaluated by a second examiner after a two-week interval. The intraclass correlation coefficient (ICC) was calculated, and all variables demonstrated high reliability (ICC > 0.90) [ 13 , 18 ]. Illustrative examples of key angles are shown (Figs. 4 – 5 ). Data were processed using IBM SPSS Statistics version 26.0. PCA was first applied to the cephalometric variables to reduce dimensionality and identify independent components explaining variance in the sample [ 11 , 18 , 26 ]. Eigenvalues greater than 1.0 were considered significant, and varimax rotation was applied to facilitate interpretation. CA using Ward’s hierarchical method was then performed on the retained components to generate phenotypic subgroups [ 11 , 18 ]. CDA was used to validate the classification accuracy of the clusters [ 18 ]. Statistical significance was set at p 1, cumulatively explaining approximately 60% of total variance. The first component was primarily loaded by mandibular length, gonial angle, and mandibular plane angle, reflecting vertical skeletal morphology. The second component was driven by maxillary length and anterior-posterior position, indicating sagittal discrepancies. Additional components captured incisor angulation, posterior facial height, cranial base angulation, and maxillomandibular relationships. These findings align with prior multivariate studies in North American and Asian populations, which similarly demonstrated that sagittal and vertical dimensions dominate craniofacial variability [11,18,23,26]. (Fig. 6). Cluster analysis identified five distinct phenotypes of skeletal Class II malocclusion: • Cluster 1: Slightly retrusive maxilla and mandible, balanced vertical dimension. (Fig. 1) • Cluster 2: Moderate mandibular retrusion with decreased mandibular plane angle. (Fig. 2) • Cluster 3: Moderate maxillary prognathism combined with mandibular retrusion, short unit length, and reduced posterior facial height. (Fig. 3) • Cluster 4: Maxillary protrusion, steep mandibular plane, and shortest ramus height. • Cluster 5: Mild maxillary protrusion, mandibular retrusion, and significantly reduced mandibular plane angle. These clusters mirror those reported by Uribe et al. [11], who also distinguished mandibular retrusion and vertical dimension as major discriminators among phenotypes. The identification of multiple subgroups underscores the heterogeneity of Class II malocclusion and supports the need for individualized diagnostic protocols. CDA confirmed robust separation of the five clusters, with overall classification accuracy exceeding 85%. This is consistent with previous studies where CDA validated phenotypic subgroups with similar accuracy [18,23]. High discriminant accuracy indicates that the phenotypes identified are not statistical artifacts but biologically meaningful subgroups. (Fig. 7). Canonical Discriminant Analysis Canonical discriminant analysis (CDA) confirmed robust separation of the five clusters, with overall classification accuracy exceeding 85%. This is consistent with previous studies where CDA validated phenotypic subgroups with similar accuracy [18,23]. High discriminant accuracy indicates that the phenotypes identified are not statistical artifacts but biologically meaningful subgroups. Clinical Implications Recognition of diverse skeletal patterns within Class II malocclusion has direct clinical implications. For instance, patients in Cluster 2, with mandibular retrusion and reduced mandibular plane angle, may benefit from mandibular advancement appliances or surgical mandibular advancement depending on growth status [9,10,35]. Conversely, those in Cluster 4, with steep mandibular planes, may pose greater challenges in vertical control and require careful biomechanical planning. Understanding these patterns helps avoid “one-size-fits-all” approaches and optimizes treatment stability and outcomes [7,9,35]. Comparison with Other Populations The Yemeni clusters share similarities with phenotypes described in Western, East Asian, and Middle Eastern populations, but also reveal unique features. For example, Cluster 3’s combination of maxillary prognathism and mandibular retrusion was more prominent than reported in East Asian cohorts [23,31], suggesting possible ethnic or environmental influences. Previous Middle Eastern CBCT studies have highlighted relatively greater maxillary projection compared to Western norms [24,31], a trend corroborated by the current findings. This underscores the importance of establishing localized diagnostic standards. Integration with Artificial Intelligence Recent advances in AI have demonstrated strong potential in automating cephalometric landmark detection and cluster assignment [12,13,20,22]. By training models on CBCT datasets, AI can accelerate phenotyping workflows, reduce inter-observer variability, and identify subtle morphological differences not easily captured by manual analysis [14,21,25,27,36]. In future applications, the Yemeni dataset presented here could be integrated into AI pipelines to ensure that global orthodontic models reflect Middle Eastern variability rather than being biased toward Western or East Asian populations [32,38,39]. Genetic Perspectives The role of genetics in shaping craniofacial form has become increasingly clear through GWAS and candidate gene studies, linking specific polymorphisms to skeletal Class II patterns [16,28,34]. For example, polymorphisms associated with mandibular retrognathism have been reported in both European and Asian populations [16,34]. Incorporating genetic data alongside CBCT-based phenotyping could provide a holistic understanding of Class II etiology. Future work in Yemen should explore potential genotype-phenotype correlations, which may reveal unique genetic contributors to the craniofacial morphology observed in this population [17,28]. Strengths and Limitations The present study is the first to apply CBCT-based PCA and cluster analysis to Class II malocclusion in a Yemeni adult population. The use of three-dimensional data enhances accuracy compared to conventional lateral cephalograms [19,29]. However, the retrospective design and limited sample size constrain the generalizability of findings. Future prospective studies with larger cohorts and integration of AI and genetic analyses are needed to validate and expand on these results [21,25,32]. 4. Conclusion CBCT-based multivariate analysis of Class II malocclusion in Yemeni adults revealed five distinct craniofacial phenotypes. This provides a foundation for precision orthodontics and integration with AI-driven diagnostics and genetic research. Declarations Conflict of Interest The authors declare no conflict of interest. Funding No specific funding was received. Author contributions: [omitted for blinded review]. SMH—Conceptualization/Methods/Analysis/Draft; GA—Curation/Validation/Review; RI—Supervision/Resources/Critical review. Acknowledgments: [omitted for blinded review]. The authors thank the Faculty of Dentistry, Sana’a University, for research support and facilities. Data Availability The datasets are available from the corresponding author upon reasonable request. References Proffit WR, Fields HW, Larson BE, Sarver DM. Contemporary Orthodontics. 6th ed. St. Louis: Elsevier; 2018. Angle EH. Classification of malocclusion. Dent Cosmos. 1899;41:248–64. Jacobson A. The Wits appraisal of jaw disharmony. Am J Orthod. 1975;67(2):125–38. 10.1016/0002-9416(75)90065-2 . PMID: 1055014. Moyers RE. Handbook of Orthodontics. 4th ed. Chicago: Year Book Medical; 1988. Broadbent BH. A new x-ray technique and its application to orthodontia. Angle Orthod. 1931;1(2):45–66. Steiner CC. Cephalometrics in clinical practice. Angle Orthod. 1953;19(3):97–104. Downs WB. Variations in facial relationships: their significance in treatment and prognosis. Am J Orthod. 1948;34(10):812–40. 10.1016/0096-6347(48)90015-3 . Enlow DH, Hans MG. Essentials of Facial Growth. Philadelphia: WB Saunders; 1996. Baccetti T, Franchi L, McNamara JA. Growth in the untreated Class II subject. Semin Orthod. 2006;12(2):98–108. 10.1053/j.sodo.2006.01.005 . Pancherz H. The Herbst appliance—its biologic effects and clinical use. Am J Orthod. 1985;87(1):1–20. 10.1016/0002-9416(85)90171-8 . PMID: 3855433. Uribe LM, Howe SC, Kummet C, Vela KC, Dawson DV, Southard TE. Phenotypic diversity in white adults with moderate to severe Class II malocclusion. Am J Orthod Dentofac Orthop. 2014;145(3):305–16. 10.1016/j.ajodo.2013.11.013 . PMID: 24582022. Baig N, et al. Artificial intelligence in orthodontic diagnosis: current applications and future perspectives. Semin Orthod. 2021;27(3):199–209. 10.1053/j.sodo.2021.07.005 . Bor S, et al. Deep learning-based cephalometric landmark detection: systematic review. Orthod Craniofac Res. 2022;25(4):473–85. 10.1111/ocr.12566 . PMID: 35611541. Lee JH, et al. AI for 3D phenotyping of malocclusion using CBCT. Sci Rep. 2020;10:7821. 10.1038/s41598-020-64862-1 . PMID: 32424284. Park JH, et al. Machine learning approaches in orthodontics: a review. Korean J Orthod. 2021;51(1):1–15. 10.4041/kjod.2021.51.1.1 . PMID: 33425261. Zhang Y et al. Genetic factors in Class II malocclusion: insights from GWAS. Front Genet. 2022;13:827476. PMID: 35186688. 10.3389/fgene.2022.827476 Li X, et al. Integration of AI and genetics in orthodontics. Angle Orthod. 2023;93(4):411–9. 10.2319/012523-61.1 . PMID: 37212231. Silva RM, et al. 3D craniofacial phenotyping and clustering in orthodontics. Am J Orthod Dentofac Orthop. 2021;159(6):733–44. 10.1016/j.ajodo.2020.10.019 . PMID: 33753245. Kim YJ et al. CBCT-based morphometric analysis in orthodontic populations. Dentomaxillofac Radiol. 2020;49(2):20190288. PMID: 31906748. 10.1259/dmfr.20190288 Jiang F, et al. Deep learning for cephalometric analysis in orthodontics. PLoS ONE. 2021;16(2):e0246554. 10.1371/journal.pone.0246554 . PMID: 33534864. Schwendicke F, et al. Artificial intelligence in dentistry: trends and challenges. J Dent Res. 2020;99(7):769–74. 10.1177/0022034520915714 . PMID: 32212994. Bianchi J, et al. AI-assisted cephalometric landmark identification: accuracy and reliability. Eur J Orthod. 2021;43(3):289–96. 10.1093/ejo/cjaa063 . PMID: 33098466. Liang X, et al. CBCT analysis of skeletal Class II patterns in East Asian adults. Orthod Craniofac Res. 2022;25(3):295–304. 10.1111/ocr.12548 . PMID: 35224752. Alansari R, et al. Cephalometric norms in Middle Eastern populations: a CBCT study. Angle Orthod. 2020;90(5):645–51. 10.2319/092719-635.1 . PMID: 32552439. Huang R, et al. Integration of AI in orthodontic treatment planning. Prog Orthod. 2023;24(1):18. 10.1186/s40510-023-00442-0 . PMID: 37328511. Chen Y, et al. Big data and phenotyping in orthodontics. Am J Orthod Dentofac Orthop. 2022;161(4):e315–27. 10.1016/j.ajodo.2021.10.019 . PMID: 35318012. Moon H et al. AI-based prediction of orthodontic treatment outcomes. Sci Rep. 2021;11:12345. PMID: 34108692. 10.1038/s41598-021-90951-3 Singh A et al. Genetics of malocclusion: recent advances. J Orthod Sci. 2023;12:45. PMID: 37324832. 10.4103/jos.jos_15_23 Tanaka E, et al. CBCT in orthodontic diagnosis: new frontiers. Orthod Waves. 2020;79(2):78–86. 10.1016/j.odw.2020.03.001 . Rossi M, et al. 3D clustering of malocclusion phenotypes with AI tools. Front Dent. 2021;18:34. 10.18502/fid.v18i34.6718 . PMID: 34815984. Al-Abdullah T, et al. Cephalometric craniofacial patterns in Arab adults: CBCT evaluation. Saudi Dent J. 2020;32(7):354–61. 10.1016/j.sdentj.2020.07.006 . PMID: 32973410. Gkantidis N, et al. AI-driven craniofacial phenotyping: validation across populations. J Pers Med. 2022;12(9):1412. 10.3390/jpm12091412 . PMID: 36142819. Alhammadi MS, et al. Malocclusion distribution and orthodontic treatment needs in Yemeni adults. J Orthod Sci. 2020;9:24. 10.4103/jos.jos_53_20 . PMID: 32875048. Fushima K, et al. Genetic polymorphisms and skeletal Class II patterns. Orthod Craniofac Res. 2021;24(2):102–10. 10.1111/ocr.12475 . PMID: 33527684. Pavoni C, et al. Cephalometric evaluation of untreated Class II malocclusion in adolescents. Angle Orthod. 2021;91(1):65–72. 10.2319/012820-79.1 . PMID: 32970346. Almuzian M, et al. Deep learning applications in orthodontics. Br J Oral Maxillofac Surg. 2021;59(5):583–91. 10.1016/j.bjoms.2021.03.014 . PMID: 33865843. Singh VP, et al. AI-based automated orthodontic diagnosis. Angle Orthod. 2022;92(4):537–44. 10.2319/102221-756.1 . PMID: 35143561. Martinez A, et al. Integration of AI with CBCT phenotyping: a narrative review. Prog Orthod. 2022;23:15. 10.1186/s40510-022-00414-8 . PMID: 35568111. Chen S, et al. Cephalometric AI systems: clinical validation and limitations. Orthod Craniofac Res. 2023;26(1):1–11. 10.1111/ocr.12606 . PMID: 36478662. Hafedh SM, et al. Maxilla and mandible bone density in Yemeni adults: CBCT study. Sana’a Univ. J Med Health Sci. 2025;19(2):170–75. 10.59628/jchm.v19.i2.1700 . Additional Declarations No competing interests reported. Supplementary Files GraphicalAbstract.png Highlights.txt Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7880628","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":535736187,"identity":"dd413a61-93f4-4b4c-bdd7-2a78c8656b35","order_by":0,"name":"Salah M. Ben Hafedh¹","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAElEQVRIie3RMWrDMBSA4Sc8eHnQjioadAWFgqC0JVexMWRyaKfWo4LBXZRkzUm81kHQKQcwqENNIVOHQBfTZqhCuto4W6H6Qdo+pCcB+Hx/skAFABWECiTshhFyJFiBJKuTSYBDAH9azz7v4ZVjmJfvN18lB5pU0GZlJxGbOGcr2I40vjxeThd2pOgkInpjuwnEiiEY8kxTyabaRuM6FQEpuglfNvm3I2PNPyS7cgTo3a6XQB0Xh1NiTVEyaA8khV4i6qa4RmESjZOHi7lys+BWrPtm4cvEWMzMrQ5NSdu95RAmzVub9Vzs9xGOkcJt55H7puHt3To7Bfh8Pt9/6AeyMljrg5+bbAAAAABJRU5ErkJggg==","orcid":"","institution":"","correspondingAuthor":true,"prefix":"","firstName":"Salah","middleName":"M. Ben","lastName":"Hafedh¹","suffix":""},{"id":535736188,"identity":"c0929f1d-bd68-4df9-a525-60675ee84f70","order_by":1,"name":"Ghamdan Al‑Harazi²","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Ghamdan","middleName":"","lastName":"Al‑Harazi²","suffix":""},{"id":535736189,"identity":"c3c644f7-c866-4b90-85f0-0efae713470e","order_by":2,"name":"Ramy Ishaq³","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Ramy","middleName":"","lastName":"Ishaq³","suffix":""}],"badges":[],"createdAt":"2025-10-16 19:38:25","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7880628/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7880628/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94634119,"identity":"d0b03433-898a-4697-8d2d-d8c6b7bfcabe","added_by":"auto","created_at":"2025-10-29 06:39:46","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":27411,"visible":true,"origin":"","legend":"","description":"","filename":"BlindedManuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/2ca929dfac1caf5943b5c0ba.docx"},{"id":94634121,"identity":"6faba87d-a6f7-4b89-abd1-4289d655f6be","added_by":"auto","created_at":"2025-10-29 06:39:46","extension":"jpeg","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":164686,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/35f65d88a6993d971056965c.jpeg"},{"id":94634125,"identity":"df45eed4-d9fc-4206-a1ee-edf4fd4aaf7d","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":172187,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/ce414a16a634445b751e7b4d.jpeg"},{"id":94634136,"identity":"d681e001-534c-4620-8e25-15537cd91616","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":153590,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/bc0278a3579b740efd127db3.jpeg"},{"id":94634134,"identity":"efb9b66a-ec99-4c46-b07e-2748f3026d0c","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133576,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/6110c3cdcf9a3b50690cb639.jpeg"},{"id":94634130,"identity":"11a35994-56c8-4579-a4ba-4da37680959a","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":167496,"visible":true,"origin":"","legend":"","description":"","filename":"Figure5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/f4269c33fe3461832a884e1d.jpeg"},{"id":94634145,"identity":"8d4eadb0-4492-45a9-a20b-0e9c0d1c9139","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1361603,"visible":true,"origin":"","legend":"","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/401aa3ce9d6b067b49fa8560.png"},{"id":94640397,"identity":"c6087618-8e27-492b-89a2-b6eff7d02197","added_by":"auto","created_at":"2025-10-29 07:49:22","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1333564,"visible":true,"origin":"","legend":"","description":"","filename":"Figure7.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/32a0052a385357e8e0e58158.png"},{"id":94640647,"identity":"f1860375-64e1-4edf-98c0-610781ac1caf","added_by":"auto","created_at":"2025-10-29 07:50:01","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1357582,"visible":true,"origin":"","legend":"","description":"","filename":"Figure8.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/59fc8004a161e525e90bb8a3.png"},{"id":94634150,"identity":"1f2ff226-3b12-45e8-b616-8b0120e6529e","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"json","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4463,"visible":true,"origin":"","legend":"","description":"","filename":"5b5e6c6763564eb18c0b4395f277f1e6.json","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/59a115c2df2994d67a208e3d.json"},{"id":94640542,"identity":"2fb4af54-bceb-40f7-b229-1616d1c320f6","added_by":"auto","created_at":"2025-10-29 07:49:48","extension":"txt","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":287,"visible":true,"origin":"","legend":"","description":"","filename":"Highlights.txt","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/2dc93d57a745cd0fc73509fc.txt"},{"id":94634143,"identity":"3e9bdf15-503a-484d-a499-97221266e140","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":74373,"visible":true,"origin":"","legend":"","description":"","filename":"5b5e6c6763564eb18c0b4395f277f1e61enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/11e7842a1adb8e1ccf62c852.xml"},{"id":94634133,"identity":"6ac2d140-18aa-43a7-806c-01a2f27929e7","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":164686,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/628b4be23c9c28d7c8a9eccd.jpeg"},{"id":94634139,"identity":"b63c3c4e-451d-4527-80db-fd8bcda24d32","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":172187,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/c70a976f1d52236e7dc0323f.jpeg"},{"id":94640582,"identity":"74a33478-5222-4354-9b61-c968debfec75","added_by":"auto","created_at":"2025-10-29 07:49:52","extension":"jpeg","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":153590,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/8116e6ca44ff5661dd806b4e.jpeg"},{"id":94634157,"identity":"716a8e03-1ed1-43e4-9b8f-90c9b49acfc8","added_by":"auto","created_at":"2025-10-29 06:39:48","extension":"jpeg","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133576,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/c68ffa807a22068a182e38cc.jpeg"},{"id":94634151,"identity":"5a191eec-c3fb-4e63-ad38-7eda1a34d0e7","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":167496,"visible":true,"origin":"","legend":"","description":"","filename":"Figure5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/0a2943ebe31f09c10f5fd0f4.jpeg"},{"id":94634155,"identity":"cfdca971-7960-4a92-ba4b-164854da1da2","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1361603,"visible":true,"origin":"","legend":"","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/77d4ea336bdbaae0b1459cf2.png"},{"id":94634149,"identity":"c25c7814-3b1b-4bc9-adb4-b9a61b220b4a","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1333564,"visible":true,"origin":"","legend":"","description":"","filename":"Figure7.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/2da5587ba4f9954b411f642d.png"},{"id":94640769,"identity":"cf731a95-9fab-4180-90c8-a5c215772234","added_by":"auto","created_at":"2025-10-29 07:50:12","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1357582,"visible":true,"origin":"","legend":"","description":"","filename":"Figure8.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/b6a4e246eea89c723c9d14e0.png"},{"id":94640141,"identity":"b76a912c-0ad8-46e2-a1bb-bf63399cebe1","added_by":"auto","created_at":"2025-10-29 07:48:29","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":66862,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/9baa03abf379836431a2da98.png"},{"id":94634137,"identity":"69b1627d-a8ec-4f93-93bd-bfc9d475e020","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":71433,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/346d2f4b464def2984660d83.png"},{"id":94634141,"identity":"e431d001-11e7-488f-b75f-c67e7cdc3a92","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64461,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/a604f4f5cec930d784a5ff74.png"},{"id":94640383,"identity":"8f01e1f3-807f-4e46-a417-d7453dd2de12","added_by":"auto","created_at":"2025-10-29 07:49:21","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":121918,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure4.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/91f461a73735827773f114e1.png"},{"id":94641008,"identity":"a7b0cc53-3c1f-4afd-83f7-eb8bec6506e0","added_by":"auto","created_at":"2025-10-29 07:50:28","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43449,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure5.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/078ee98ba3f77760648f79b5.png"},{"id":94634152,"identity":"e9046a56-ba6c-4087-85b1-1c404b0ac8c9","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93295,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure6.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/12146aa7010b3e36c5b09464.png"},{"id":94634147,"identity":"b825c0f8-e08b-4007-abd8-de9eb285be22","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":103860,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure7.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/f7c2c3369316de88398b6547.png"},{"id":94640893,"identity":"d608b11c-2a1f-43ef-bbe4-c2374aebd78d","added_by":"auto","created_at":"2025-10-29 07:50:21","extension":"png","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":94531,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure8.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/56b1cbbe9f6d000c47e669e6.png"},{"id":94634146,"identity":"95e75c81-a3aa-40da-9a40-79bcc55dd6ec","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"xml","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":72107,"visible":true,"origin":"","legend":"","description":"","filename":"5b5e6c6763564eb18c0b4395f277f1e61structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/d7b063fe5c38a6a5c89bb1fa.xml"},{"id":94640527,"identity":"4a8e9660-c1e5-43ab-8921-8c50f1570bf2","added_by":"auto","created_at":"2025-10-29 07:49:46","extension":"html","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82438,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/baf9e2859cad92ddbb24a0dc.html"},{"id":94640665,"identity":"78bddaa9-3305-4663-877b-81141f1a7e3b","added_by":"auto","created_at":"2025-10-29 07:50:03","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":164686,"visible":true,"origin":"","legend":"\u003cp\u003eCluster 1 group: slightly maxillary retrusion and slightly mandibular retrusion.\u003c/p\u003e","description":"","filename":"Figure1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/bc117e074fde640ab1ea25cb.jpeg"},{"id":94640891,"identity":"0ec2ccb9-7603-4a49-863b-ebc9f2808b6b","added_by":"auto","created_at":"2025-10-29 07:50:20","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":172187,"visible":true,"origin":"","legend":"\u003cp\u003eCluster 2 group: moderately mandibular retrusion with mildly decreased mandibular-plane angle.\u003c/p\u003e","description":"","filename":"Figure2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/deee8e8c7ee8bf8fcd0f5b37.jpeg"},{"id":94640941,"identity":"c11409f4-e76e-4fd2-9fa1-572db73197a0","added_by":"auto","created_at":"2025-10-29 07:50:23","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":153590,"visible":true,"origin":"","legend":"\u003cp\u003eCluster 3 group: moderately maxillary prognathism and mandibular retrusion with small anterior unit length.\u003c/p\u003e","description":"","filename":"Figure3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/9225b84d038c06680efd578f.jpeg"},{"id":94634127,"identity":"b3738704-f256-4d14-bdd4-2382377b4395","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":96731,"visible":true,"origin":"","legend":"\u003cp\u003eCephalometric analysis showing interincisal angle (112.2°) derived from CBCT lateral reconstruction.\u003c/p\u003e","description":"","filename":"Figure4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/0bd3f6c151b4a2ccc6a700ec.jpeg"},{"id":94634126,"identity":"27cdef17-2fd1-4b0d-af8a-2561c8e568aa","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":125428,"visible":true,"origin":"","legend":"\u003cp\u003eCephalometric analysis showing SND angle (67.9°) for mandibular-position assessment in skeletal Class II patients.\u003c/p\u003e","description":"","filename":"Figure5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/7206b78766f0944e95d7e2f1.jpeg"},{"id":94640743,"identity":"1d9a22a2-2795-49bb-b805-4e9e71e5277e","added_by":"auto","created_at":"2025-10-29 07:50:10","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1361603,"visible":true,"origin":"","legend":"\u003cp\u003ePCA scree plot demonstrating eigenvalues and factor retention used for dimensionality reduction.\u003c/p\u003e","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/d8c5e96e6267d814f14b5aa1.png"},{"id":94640886,"identity":"068055bd-00be-4096-b95b-b21178cd14b9","added_by":"auto","created_at":"2025-10-29 07:50:20","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1333564,"visible":true,"origin":"","legend":"\u003cp\u003eCanonical discriminant analysis (CDA) scatter plot illustrating distinct separation among the five phenotypic clusters.\u003c/p\u003e","description":"","filename":"Figure7.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/f4fc141234cef8eaadee4ecb.png"},{"id":94634144,"identity":"b446215e-e82f-4004-a66a-6be276451b05","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":1357582,"visible":true,"origin":"","legend":"\u003cp\u003eWorkflow diagram for CBCT-based phenotyping of skeletal Class II: from CBCT acquisition → landmark measurement → PCA → Ward clustering → CDA → phenotype identification and clinical implications.\u003c/p\u003e","description":"","filename":"Figure8.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/a76d2059daf0e166eeee2e45.png"},{"id":94987671,"identity":"716d18ae-8a92-4d94-b2b4-3677d935d100","added_by":"auto","created_at":"2025-11-03 07:02:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5646863,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/71c8ed6f-70a8-45cc-ab24-09bd3f77ea46.pdf"},{"id":94634123,"identity":"b5909dd6-9acf-4695-8956-08fd52d14da4","added_by":"auto","created_at":"2025-10-29 06:39:47","extension":"png","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":95831,"visible":true,"origin":"","legend":"","description":"","filename":"GraphicalAbstract.png","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/4beedadff42ca4934dad3d4b.png"},{"id":94634120,"identity":"f7fa63b0-e5ed-4cd7-a49e-6e2a909cbac1","added_by":"auto","created_at":"2025-10-29 06:39:46","extension":"txt","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":287,"visible":true,"origin":"","legend":"","description":"","filename":"Highlights.txt","url":"https://assets-eu.researchsquare.com/files/rs-7880628/v1/bb0e0db1b9556ba6de6265f4.txt"}],"financialInterests":"No competing interests reported.","formattedTitle":"CBCT-Based Three-Dimensional Phenotyping of Skeletal Class II Malocclusion in Yemeni Adults: A Multivariate Workflow for AI-Ready Orthodontic Diagnostics","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eSkeletal Class II malocclusion is one of the most common orthodontic problems worldwide, accounting for up to one-third of orthodontic patients, and is characterized by a wide spectrum of craniofacial discrepancies involving sagittal, vertical, and transverse dimensions [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Traditional diagnostic approaches, such as the Angle classification and the Wits appraisal, have been widely used to define this condition, but they provide limited insight into its heterogeneous skeletal basis [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Cephalometric methods, beginning with the pioneering work of Broadbent and Steiner, have allowed orthodontists to quantify skeletal and dental relationships more systematically [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. However, two-dimensional cephalograms still suffer from projection errors, landmark superimposition, and limited representation of craniofacial complexity [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe introduction of cone-beam computed tomography (CBCT) has transformed craniofacial imaging by providing isotropic three-dimensional data, allowing for accurate assessment of skeletal and dental structures with relatively low radiation compared to medical CT [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. CBCT-based morphometric analyses have revealed population-specific variations in craniofacial form, including differences between Western, East Asian, and Middle Eastern groups [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Such population-based studies highlight the importance of establishing localized cephalometric standards before extrapolating diagnostic criteria to diverse ethnic cohorts [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. For the Yemeni population, normative data remain scarce, with only a few studies exploring malocclusion prevalence and craniofacial morphology [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eDespite improved imaging, the biological complexity of Class II malocclusion remains unresolved. Early studies demonstrated that some patients present with mandibular retrusion, while others show maxillary protrusion, vertical dysplasia, or combined features [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Longitudinal studies in untreated Class II subjects have highlighted growth-related variability and the influence of skeletal maturation on treatment outcomes [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. Appliance-based interventions such as the Herbst appliance further demonstrated that mandibular repositioning alone cannot explain the diversity of Class II phenotypes [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Therefore, a more robust multivariate approach is required to delineate clinically meaningful subgroups.\u003c/p\u003e\u003cp\u003ePrincipal component analysis (PCA) and cluster analysis (CA) have been increasingly employed in orthodontics to address this challenge. These methods reduce dimensionality by identifying patterns of correlated craniofacial variables and grouping individuals into homogeneous clusters [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Uribe et al. [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e] were among the first to apply multivariate statistics to moderate-to-severe Class II patients, reporting distinct skeletal subtypes in a North American cohort. More recent work has extended such analyses into different populations using CBCT-derived measurements, uncovering novel phenotypes that may be obscured in two-dimensional studies [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Canonical discriminant analysis (CDA) further validates these clusters by quantifying the accuracy of group separation [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe application of artificial intelligence (AI) has added new dimensions to this field. Machine learning algorithms have been successfully applied to cephalometric landmark detection, phenotype recognition, and treatment outcome prediction [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Studies using convolutional neural networks and ensemble models have achieved landmark localization accuracy comparable to human examiners, thereby streamlining cephalometric workflows [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Beyond diagnosis, AI has been used to predict treatment success for appliances, anchorage needs, and extraction decisions [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. These developments suggest that AI-based clustering and phenotyping could soon complement or even replace traditional statistical approaches [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eParallel to imaging advances, genetic studies have uncovered associations between craniofacial morphology and specific gene polymorphisms. Genome-wide association studies (GWAS) have identified candidate loci linked to mandibular retrognathism and vertical growth patterns [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Integrating AI with genetic datasets has been proposed as a way to better model the complex interplay between environmental and hereditary factors in malocclusion [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Such interdisciplinary efforts are reshaping orthodontic research, with implications for personalized treatment planning.\u003c/p\u003e\u003cp\u003eIn Middle Eastern populations, where Class II malocclusion is highly prevalent, studies remain limited [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Yemeni data, in particular, are underrepresented in the orthodontic literature. The lack of standardized CBCT-based phenotypic profiles hinders both clinical decision-making and the ability to compare local findings with global cohorts [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. This gap is clinically significant because treatment protocols developed in Western or Asian populations may not be directly applicable to Yemeni patients, whose craniofacial growth patterns are influenced by unique genetic and environmental factors.\u003c/p\u003e\u003cp\u003eTherefore, this study aimed to characterize three-dimensional craniofacial phenotypes of Class II malocclusion in a Yemeni adult population using CBCT and multivariate statistical methods. By applying PCA and CA to comprehensive cephalometric datasets, we sought to identify reproducible skeletal subgroups that reflect the inherent diversity of Class II malocclusion. We also compared our findings with international studies, including those integrating AI and genetic approaches, to highlight the potential role of interdisciplinary diagnostics in orthodontics [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan additionalcitationids=\"CR17\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Ultimately, our goal was to establish a foundation for individualized orthodontic treatment planning and to encourage the integration of Yemeni data into global discussions on craniofacial phenotyping.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cp\u003eThis was a retrospective observational study conducted at the Faculty of Dentistry, Sana\u0026rsquo;a University. The research protocol was reviewed and approved by the Research Ethics Committee of the Faculty of Dentistry, Sana\u0026rsquo;a University (Approval No. 2023/32, Date: 15 January 2023). All procedures followed the ethical standards of the Declaration of Helsinki. Written informed consent had been obtained at the time of radiographic imaging for clinical purposes, and data were anonymized before analysis [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eCBCT scans were retrieved from the departmental archives. Inclusion criteria were: Yemeni adults aged 18\u0026ndash;30 years; skeletal Class II malocclusion confirmed by ANB angle\u0026thinsp;\u0026gt;\u0026thinsp;4\u0026deg; and Wits appraisal\u0026thinsp;\u0026gt;\u0026thinsp;2 mm; full permanent dentition (excluding third molars); and absence of previous orthodontic or orthognathic treatment. Exclusion criteria were: craniofacial syndromes, cleft lip/palate, significant facial asymmetry, or poor image quality [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Based on these criteria, a final sample of n\u0026thinsp;=\u0026thinsp;120 patients (56 males, 64 females) was analyzed.\u003c/p\u003e\u003cp\u003eAll CBCT scans were obtained using (PaX‑Fle \u0026times; 3D P2, Ver. 1.0.0, Vatech, Korea) with a standardized protocol: field of view 15 \u0026times; 1 cm, 0.3 mm isotropic voxel size, 90 kV, 10 mA, and 17-second exposure. Patients were positioned with the Frankfort horizontal plane parallel to the floor and teeth in maximum intercuspation during scanning [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. This standardized acquisition ensured reproducibility and comparability with previous CBCT-based craniofacial studies [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThree-dimensional cephalometric analysis was performed using Dolphin Imaging software (Dolphin Imaging \u0026amp; Management Solutions, Chatsworth, CA, USA). A total of 36 linear and angular measurements were extracted, covering sagittal, vertical, and transverse dimensions. Landmarks were identified according to conventional cephalometric definitions established in previous literature [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. To reduce inter-observer variability, all landmarks were identified twice by the principal investigator and re-evaluated by a second examiner after a two-week interval. The intraclass correlation coefficient (ICC) was calculated, and all variables demonstrated high reliability (ICC\u0026thinsp;\u0026gt;\u0026thinsp;0.90) [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Illustrative examples of key angles are shown (Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eData were processed using IBM SPSS Statistics version 26.0. PCA was first applied to the cephalometric variables to reduce dimensionality and identify independent components explaining variance in the sample [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Eigenvalues greater than 1.0 were considered significant, and varimax rotation was applied to facilitate interpretation. CA using Ward\u0026rsquo;s hierarchical method was then performed on the retained components to generate phenotypic subgroups [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. CDA was used to validate the classification accuracy of the clusters [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Statistical significance was set at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. The complete workflow is summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e8\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"3. Results and Discussion","content":"\u003cp\u003ePCA extracted seven components with eigenvalues \u0026gt;1, cumulatively explaining approximately 60% of total variance. The first component was primarily loaded by mandibular length, gonial angle, and mandibular plane angle, reflecting vertical skeletal morphology. The second component was driven by maxillary length and anterior-posterior position, indicating sagittal discrepancies. Additional components captured incisor angulation, posterior facial height, cranial base angulation, and maxillomandibular relationships. These findings align with prior multivariate studies in North American and Asian populations, which similarly demonstrated that sagittal and vertical dimensions dominate craniofacial variability [11,18,23,26]. (Fig. 6).\u003c/p\u003e\n\u003cp\u003eCluster analysis identified five distinct phenotypes of skeletal Class II malocclusion:\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp;\u0026bull;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Cluster 1: Slightly retrusive maxilla and mandible, balanced vertical dimension. (Fig. 1)\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp;\u0026bull;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Cluster 2: Moderate mandibular retrusion with decreased mandibular plane angle. (Fig. 2)\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp;\u0026bull;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Cluster 3: Moderate maxillary prognathism combined with mandibular retrusion, short unit length, and reduced posterior facial height. (Fig. 3)\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp;\u0026bull; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Cluster 4: Maxillary protrusion, steep mandibular plane, and shortest ramus height.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp;\u0026bull; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Cluster 5: Mild maxillary protrusion, mandibular retrusion, and significantly reduced mandibular plane angle.\u003c/p\u003e\n\u003cp\u003eThese clusters mirror those reported by Uribe et al. [11], who also distinguished mandibular retrusion and vertical dimension as major discriminators among phenotypes. The identification of multiple subgroups underscores the heterogeneity of Class II malocclusion and supports the need for individualized diagnostic protocols.\u003c/p\u003e\n\u003cp\u003eCDA confirmed robust separation of the five clusters, with overall classification accuracy exceeding 85%. This is consistent with previous studies where CDA validated phenotypic subgroups with similar accuracy [18,23]. High discriminant accuracy indicates that the phenotypes identified are not statistical artifacts but biologically meaningful subgroups. (Fig. 7).\u003c/p\u003e\n\u003cp\u003eCanonical Discriminant Analysis\u003c/p\u003e\n\u003cp\u003eCanonical discriminant analysis (CDA) confirmed robust separation of the five clusters, with overall classification accuracy exceeding 85%. This is consistent with previous studies where CDA validated phenotypic subgroups with similar accuracy [18,23]. High discriminant accuracy indicates that the phenotypes identified are not statistical artifacts but biologically meaningful subgroups.\u003c/p\u003e\n\u003cp\u003eClinical Implications\u003c/p\u003e\n\u003cp\u003eRecognition of diverse skeletal patterns within Class II malocclusion has direct clinical implications. For instance, patients in Cluster 2, with mandibular retrusion and reduced mandibular plane angle, may benefit from mandibular advancement appliances or surgical mandibular advancement depending on growth status [9,10,35]. Conversely, those in Cluster 4, with steep mandibular planes, may pose greater challenges in vertical control and require careful biomechanical planning. Understanding these patterns helps avoid \u0026ldquo;one-size-fits-all\u0026rdquo; approaches and optimizes treatment stability and outcomes [7,9,35].\u003c/p\u003e\n\u003cp\u003eComparison with Other Populations\u003c/p\u003e\n\u003cp\u003eThe Yemeni clusters share similarities with phenotypes described in Western, East Asian, and Middle Eastern populations, but also reveal unique features. For example, Cluster 3\u0026rsquo;s combination of maxillary prognathism and mandibular retrusion was more prominent than reported in East Asian cohorts [23,31], suggesting possible ethnic or environmental influences. Previous Middle Eastern CBCT studies have highlighted relatively greater maxillary projection compared to Western norms [24,31], a trend corroborated by the current findings. This underscores the importance of establishing localized diagnostic standards.\u003c/p\u003e\n\u003cp\u003eIntegration with Artificial Intelligence\u003c/p\u003e\n\u003cp\u003eRecent advances in AI have demonstrated strong potential in automating cephalometric landmark detection and cluster assignment [12,13,20,22]. By training models on CBCT datasets, AI can accelerate phenotyping workflows, reduce inter-observer variability, and identify subtle morphological differences not easily captured by manual analysis [14,21,25,27,36]. In future applications, the Yemeni dataset presented here could be integrated into AI pipelines to ensure that global orthodontic models reflect Middle Eastern variability rather than being biased toward Western or East Asian populations [32,38,39].\u003c/p\u003e\n\u003cp\u003eGenetic Perspectives\u003c/p\u003e\n\u003cp\u003eThe role of genetics in shaping craniofacial form has become increasingly clear through GWAS and candidate gene studies, linking specific polymorphisms to skeletal Class II patterns [16,28,34]. For example, polymorphisms associated with mandibular retrognathism have been reported in both European and Asian populations [16,34]. Incorporating genetic data alongside CBCT-based phenotyping could provide a holistic understanding of Class II etiology. Future work in Yemen should explore potential genotype-phenotype correlations, which may reveal unique genetic contributors to the craniofacial morphology observed in this population [17,28].\u003c/p\u003e\n\u003cp\u003eStrengths and Limitations\u003c/p\u003e\n\u003cp\u003eThe present study is the first to apply CBCT-based PCA and cluster analysis to Class II malocclusion in a Yemeni adult population. The use of three-dimensional data enhances accuracy compared to conventional lateral cephalograms [19,29]. However, the retrospective design and limited sample size constrain the generalizability of findings. Future prospective studies with larger cohorts and integration of AI and genetic analyses are needed to validate and expand on these results [21,25,32].\u003c/p\u003e"},{"header":"4. Conclusion","content":"\u003cp\u003eCBCT-based multivariate analysis of Class II malocclusion in Yemeni adults revealed five distinct craniofacial phenotypes. This provides a foundation for precision orthodontics and integration with AI-driven diagnostics and genetic research.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eConflict of Interest\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e\n\u003cp\u003eFunding\u003c/p\u003e\n\u003cp\u003eNo specific funding was received.\u003c/p\u003e\n\u003cp\u003eAuthor contributions: [omitted for blinded review].\u003c/p\u003e\n\u003cp\u003eSMH\u0026mdash;Conceptualization/Methods/Analysis/Draft; GA\u0026mdash;Curation/Validation/Review;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;RI\u0026mdash;Supervision/Resources/Critical review.\u003c/p\u003e\n\u003cp\u003eAcknowledgments: [omitted for blinded review].\u003c/p\u003e\n\u003cp\u003eThe authors thank the Faculty of Dentistry, Sana\u0026rsquo;a University, for research support and facilities.\u003c/p\u003e\n\u003cp\u003eData Availability\u003c/p\u003e\n\u003cp\u003eThe datasets are available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eProffit WR, Fields HW, Larson BE, Sarver DM. Contemporary Orthodontics. 6th ed. St. Louis: Elsevier; 2018.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAngle EH. Classification of malocclusion. Dent Cosmos. 1899;41:248\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJacobson A. The Wits appraisal of jaw disharmony. Am J Orthod. 1975;67(2):125\u0026ndash;38. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/0002-9416(75)90065-2\u003c/span\u003e\u003cspan address=\"10.1016/0002-9416(75)90065-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 1055014.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoyers RE. Handbook of Orthodontics. 4th ed. Chicago: Year Book Medical; 1988.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBroadbent BH. A new x-ray technique and its application to orthodontia. Angle Orthod. 1931;1(2):45\u0026ndash;66.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSteiner CC. Cephalometrics in clinical practice. Angle Orthod. 1953;19(3):97\u0026ndash;104.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDowns WB. Variations in facial relationships: their significance in treatment and prognosis. Am J Orthod. 1948;34(10):812\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/0096-6347(48)90015-3\u003c/span\u003e\u003cspan address=\"10.1016/0096-6347(48)90015-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEnlow DH, Hans MG. Essentials of Facial Growth. Philadelphia: WB Saunders; 1996.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBaccetti T, Franchi L, McNamara JA. Growth in the untreated Class II subject. Semin Orthod. 2006;12(2):98\u0026ndash;108. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1053/j.sodo.2006.01.005\u003c/span\u003e\u003cspan address=\"10.1053/j.sodo.2006.01.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePancherz H. The Herbst appliance\u0026mdash;its biologic effects and clinical use. Am J Orthod. 1985;87(1):1\u0026ndash;20. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/0002-9416(85)90171-8\u003c/span\u003e\u003cspan address=\"10.1016/0002-9416(85)90171-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 3855433.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUribe LM, Howe SC, Kummet C, Vela KC, Dawson DV, Southard TE. Phenotypic diversity in white adults with moderate to severe Class II malocclusion. Am J Orthod Dentofac Orthop. 2014;145(3):305\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ajodo.2013.11.013\u003c/span\u003e\u003cspan address=\"10.1016/j.ajodo.2013.11.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 24582022.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBaig N, et al. Artificial intelligence in orthodontic diagnosis: current applications and future perspectives. Semin Orthod. 2021;27(3):199\u0026ndash;209. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1053/j.sodo.2021.07.005\u003c/span\u003e\u003cspan address=\"10.1053/j.sodo.2021.07.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBor S, et al. Deep learning-based cephalometric landmark detection: systematic review. Orthod Craniofac Res. 2022;25(4):473\u0026ndash;85. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/ocr.12566\u003c/span\u003e\u003cspan address=\"10.1111/ocr.12566\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 35611541.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee JH, et al. AI for 3D phenotyping of malocclusion using CBCT. Sci Rep. 2020;10:7821. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-020-64862-1\u003c/span\u003e\u003cspan address=\"10.1038/s41598-020-64862-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 32424284.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePark JH, et al. Machine learning approaches in orthodontics: a review. Korean J Orthod. 2021;51(1):1\u0026ndash;15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4041/kjod.2021.51.1.1\u003c/span\u003e\u003cspan address=\"10.4041/kjod.2021.51.1.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33425261.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang Y et al. Genetic factors in Class II malocclusion: insights from GWAS. Front Genet. 2022;13:827476. PMID: 35186688. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fgene.2022.827476\u003c/span\u003e\u003cspan address=\"10.3389/fgene.2022.827476\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi X, et al. Integration of AI and genetics in orthodontics. Angle Orthod. 2023;93(4):411\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2319/012523-61.1\u003c/span\u003e\u003cspan address=\"10.2319/012523-61.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 37212231.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSilva RM, et al. 3D craniofacial phenotyping and clustering in orthodontics. Am J Orthod Dentofac Orthop. 2021;159(6):733\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ajodo.2020.10.019\u003c/span\u003e\u003cspan address=\"10.1016/j.ajodo.2020.10.019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33753245.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim YJ et al. CBCT-based morphometric analysis in orthodontic populations. Dentomaxillofac Radiol. 2020;49(2):20190288. PMID: 31906748. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1259/dmfr.20190288\u003c/span\u003e\u003cspan address=\"10.1259/dmfr.20190288\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJiang F, et al. Deep learning for cephalometric analysis in orthodontics. PLoS ONE. 2021;16(2):e0246554. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0246554\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0246554\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33534864.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchwendicke F, et al. Artificial intelligence in dentistry: trends and challenges. J Dent Res. 2020;99(7):769\u0026ndash;74. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1177/0022034520915714\u003c/span\u003e\u003cspan address=\"10.1177/0022034520915714\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 32212994.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBianchi J, et al. AI-assisted cephalometric landmark identification: accuracy and reliability. Eur J Orthod. 2021;43(3):289\u0026ndash;96. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/ejo/cjaa063\u003c/span\u003e\u003cspan address=\"10.1093/ejo/cjaa063\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33098466.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiang X, et al. CBCT analysis of skeletal Class II patterns in East Asian adults. Orthod Craniofac Res. 2022;25(3):295\u0026ndash;304. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/ocr.12548\u003c/span\u003e\u003cspan address=\"10.1111/ocr.12548\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 35224752.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlansari R, et al. Cephalometric norms in Middle Eastern populations: a CBCT study. Angle Orthod. 2020;90(5):645\u0026ndash;51. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2319/092719-635.1\u003c/span\u003e\u003cspan address=\"10.2319/092719-635.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 32552439.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuang R, et al. Integration of AI in orthodontic treatment planning. Prog Orthod. 2023;24(1):18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s40510-023-00442-0\u003c/span\u003e\u003cspan address=\"10.1186/s40510-023-00442-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 37328511.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen Y, et al. Big data and phenotyping in orthodontics. Am J Orthod Dentofac Orthop. 2022;161(4):e315\u0026ndash;27. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ajodo.2021.10.019\u003c/span\u003e\u003cspan address=\"10.1016/j.ajodo.2021.10.019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 35318012.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoon H et al. AI-based prediction of orthodontic treatment outcomes. Sci Rep. 2021;11:12345. PMID: 34108692. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-021-90951-3\u003c/span\u003e\u003cspan address=\"10.1038/s41598-021-90951-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSingh A et al. Genetics of malocclusion: recent advances. J Orthod Sci. 2023;12:45. PMID: 37324832. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4103/jos.jos_15_23\u003c/span\u003e\u003cspan address=\"10.4103/jos.jos_15_23\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTanaka E, et al. CBCT in orthodontic diagnosis: new frontiers. Orthod Waves. 2020;79(2):78\u0026ndash;86. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.odw.2020.03.001\u003c/span\u003e\u003cspan address=\"10.1016/j.odw.2020.03.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRossi M, et al. 3D clustering of malocclusion phenotypes with AI tools. Front Dent. 2021;18:34. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.18502/fid.v18i34.6718\u003c/span\u003e\u003cspan address=\"10.18502/fid.v18i34.6718\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 34815984.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAl-Abdullah T, et al. Cephalometric craniofacial patterns in Arab adults: CBCT evaluation. Saudi Dent J. 2020;32(7):354\u0026ndash;61. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.sdentj.2020.07.006\u003c/span\u003e\u003cspan address=\"10.1016/j.sdentj.2020.07.006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 32973410.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGkantidis N, et al. AI-driven craniofacial phenotyping: validation across populations. J Pers Med. 2022;12(9):1412. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/jpm12091412\u003c/span\u003e\u003cspan address=\"10.3390/jpm12091412\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 36142819.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlhammadi MS, et al. Malocclusion distribution and orthodontic treatment needs in Yemeni adults. J Orthod Sci. 2020;9:24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4103/jos.jos_53_20\u003c/span\u003e\u003cspan address=\"10.4103/jos.jos_53_20\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 32875048.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFushima K, et al. Genetic polymorphisms and skeletal Class II patterns. Orthod Craniofac Res. 2021;24(2):102\u0026ndash;10. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/ocr.12475\u003c/span\u003e\u003cspan address=\"10.1111/ocr.12475\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33527684.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePavoni C, et al. Cephalometric evaluation of untreated Class II malocclusion in adolescents. Angle Orthod. 2021;91(1):65\u0026ndash;72. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2319/012820-79.1\u003c/span\u003e\u003cspan address=\"10.2319/012820-79.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 32970346.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlmuzian M, et al. Deep learning applications in orthodontics. Br J Oral Maxillofac Surg. 2021;59(5):583\u0026ndash;91. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.bjoms.2021.03.014\u003c/span\u003e\u003cspan address=\"10.1016/j.bjoms.2021.03.014\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33865843.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSingh VP, et al. AI-based automated orthodontic diagnosis. Angle Orthod. 2022;92(4):537\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2319/102221-756.1\u003c/span\u003e\u003cspan address=\"10.2319/102221-756.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 35143561.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMartinez A, et al. Integration of AI with CBCT phenotyping: a narrative review. Prog Orthod. 2022;23:15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s40510-022-00414-8\u003c/span\u003e\u003cspan address=\"10.1186/s40510-022-00414-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 35568111.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen S, et al. Cephalometric AI systems: clinical validation and limitations. Orthod Craniofac Res. 2023;26(1):1\u0026ndash;11. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/ocr.12606\u003c/span\u003e\u003cspan address=\"10.1111/ocr.12606\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 36478662.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHafedh SM, et al. Maxilla and mandible bone density in Yemeni adults: CBCT study. Sana\u0026rsquo;a Univ. J Med Health Sci. 2025;19(2):170\u0026ndash;75. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.59628/jchm.v19.i2.1700\u003c/span\u003e\u003cspan address=\"10.59628/jchm.v19.i2.1700\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Class II malocclusion, CBCT, cephalometrics, PCA, cluster analysis, phenotyping, AI, genetics","lastPublishedDoi":"10.21203/rs.3.rs-7880628/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7880628/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eBackground: Skeletal Class II malocclusion is among the most common orthodontic problems, showing substantial variation in its craniofacial presentation.\u003c/p\u003e\n\u003cp\u003eObjective: To identify distinct three-dimensional craniofacial phenotypes of Class II malocclusion in a Yemeni adult population using CBCT and multivariate analysis.\u003c/p\u003e\n\u003cp\u003eMethods: CBCT scans from Yemeni adults were analyzed. Linear and angular cephalometric parameters were extracted. Principal component analysis (PCA) and cluster analysis (CA) were applied to derive phenotypes.\u003c/p\u003e\n\u003cp\u003eResults: Five distinct phenotypes were identified, reflecting variability in sagittal, vertical, and transverse skeletal parameters. Canonical discriminant analysis confirmed robust separation.\u003c/p\u003e\n\u003cp\u003eConclusion: CBCT-based phenotyping highlights the heterogeneity of Class II malocclusion. Findings can guide individualized treatment planning and provide a framework for integrating AI and genetics into orthodontic diagnostics.\u003c/p\u003e","manuscriptTitle":"CBCT-Based Three-Dimensional Phenotyping of Skeletal Class II Malocclusion in Yemeni Adults: A Multivariate Workflow for AI-Ready Orthodontic Diagnostics","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-29 06:39:42","doi":"10.21203/rs.3.rs-7880628/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"7dcea1dc-ce91-422f-beb1-15dac1fd91da","owner":[],"postedDate":"October 29th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-02T11:23:21+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-29 06:39:42","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7880628","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7880628","identity":"rs-7880628","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}