Digital Orthodontic Setup as a Predictive Tool for Post-treatment Open Gingival Embrasures in Adult Extraction Cases with Fixed Appliances: A Retrospective Study

Digital Orthodontic Setup as a Predictive Tool for Post-treatment Open Gingival Embrasures in Adult Extraction Cases with Fixed Appliances: A Retrospective Study | 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 Digital Orthodontic Setup as a Predictive Tool for Post-treatment Open Gingival Embrasures in Adult Extraction Cases with Fixed Appliances: A Retrospective Study Xinyi Ren, Jiaxin Deng, Lang Lei, Jialing Li, Huang Li This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8439148/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 Objectives To evaluate whether digital orthodontic setups can predict the presence and severity of postoperative open gingival embrasures (OGEs) in adult extraction cases treated with fixed appliances. Materials and Methods This retrospective study included 62 adults treated with four-premolar extractions and fixed appliances (620 anterior interproximal sites). Actual OGEs were assessed on post-treatment intraoral photographs. Predicted OGEs were assessed on digital setups generated on a clear aligner platform. Binary performance was evaluated using accuracy, sensitivity, specificity, PPV, NPV, F1 score, and Cohen’s κ. Severity agreement (modified Jemt grades) was assessed using weighted κ, mean absolute error (MAE), and Spearman’s ρ. Analyses were performed overall and by arch, sex, and site. Results Predicted OGE incidence approximated the observed rate (76.5% vs 71.9%; bias +4.5%). Binary prediction showed high performance (accuracy 91.6%, sensitivity 0.916, specificity 0.918, PPV 97.3%, NPV 77.0%, F1 0.943, κ 0.823). Severity prediction showed substantial agreement (weighted κ 0.716–0.774; MAE 0.194; ρ 0.786), with exact matches in 80.8% of sites and ±1-grade agreement in 99.8%. Conclusions Digital setups provide clinically meaningful prediction of postoperative OGEs after fixed-appliance extraction treatment, with modest overprediction. Clinical Relevance Digital setup–based prediction enables early identification of embrasure sites at higher esthetic risk, supporting proactive soft-tissue risk assessment and informed treatment planning in adult extraction orthodontics. Open gingival embrasures Orthodontic tooth movement Digital orthodontic setup Extraction orthodontics Soft tissue esthetics. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 1. INTRODUCTION Open gingival embrasures (OGEs), commonly referred to as “black triangles”, are a frequent esthetic and functional concern following orthodontic treatment, particularly in adult patients[ 1 ]. Epidemiological studies have reported OGE prevalence ranging from approximately 20% to 38%, with higher occurrence in the anterior region with triangular tooth morphology, dental crowding, or compromised periodontal support[ 1 – 4 ]. Characterized by incomplete interdental papilla fill, OGEs negatively affect smile esthetics and may promote plaque accumulation, food impaction, and phonetic difficulties, thereby reducing periodontal stability and patient satisfaction[ 5 – 7 ]. As esthetic expectations among adult orthodontic patients continue to rise, even minor papillary deficiencies may disproportionately influence perceived treatment success[ 8 ]. Extraction therapy further increases susceptibility to OGEs because it involves greater tooth movement and alveolar bone remodeling[ 2 ]. Therefore, prediction of OGEs during treatment planning is essential to reduce esthetic risk, particularly in adult extraction cases. The development of OGEs is multifactorial and closely related to anatomical, periodontal, and biomechanical determinants[ 5 , 9 , 10 ]. Widely recognized risk factors include patient age, gingival biotype, interdental bone height, crown morphology, root angulation, and the vertical distance between the contact point and the alveolar crest[ 1 , 3 , 5 , 10 – 13 ]. Post-space-closure root divergence and reduced interdental contact area have also been associated with papillary loss[ 1 , 4 , 13 ]. Despite these established predictors, clinical risk assessment remains largely qualitative and relies on static two-dimensional records such as intraoral photographs and periapical radiographs[ 14 , 15 ]. These methods provide limited information on three-dimensional interdental geometry and cannot simulate tooth movement-related changes, leading clinicians to depend heavily on experience-based judgment. Digital orthodontic setups, originally developed for clear aligner therapy, enable accurate three-dimensional reconstruction and virtual simulation of tooth movement. Previous studies have demonstrated acceptable predictability of aligner-based setups for various tooth movements, including arch expansion, molar distalization, alignment efficiency, torque control, and intrusion[ 16 – 19 ]. However, existing research has largely focused on hard-tissue, with limited attention to soft-tissue esthetics such as interdental papilla, and setup-based prediction in fixed appliance therapy remains underexplored. A recent prospective study reported acceptable average agreement but substantial individual discrepancies between planned and achieved tooth positions in fixed appliance treatment[ 20 ]. In addition, our pilot study showed that Invisalign ClinCheck could reasonably predict postoperative OGEs treated with clear aligners[ 15 ]. Whether such predictive validity extends to fixed orthodontic treatment has not been characterized. Geometric determinants encoded within digital setups—such as contact-point position, crown morphology, and root alignment—are known to be closely associated with papilla presence[ 2 , 3 , 9 , 12 ]. Therefore, digital setups may function for estimating papillary risk, even in the absence of explicit soft-tissue simulation. Notably, our prior study showed that clear aligners were associated with a higher incidence and severity of OGEs than fixed appliances[ 21 ], underscoring the importance of appliance-specific validation. Therefore, this retrospective study evaluated the predictive validity of digital setups for postoperative OGEs in adult extraction patients treated with fixed appliances, using binary and ordinal (Jemt) outcomes with subgroup analyse. This study seeks to support the use of digital setup-based assessment in orthodontic soft-tissue esthetic risk evaluation. 2. MATERIALS AND METHODS PARTICIPANTS Patients This retrospective study was approved by the Ethics Committee of Nanjing Stomatological Hospital, Nanjing University (Approval No. NJSH-2023NL-036). Adult patients who initiated fixed orthodontic treatment between January 2020 and December 2022 at the same institution were consecutively screened (Fig.1). Sample size and power A precision-based design was used, with overall weighted Cohen’s kappa (κw) for Jemt grades (0-3) as the primary endpoint. Sample size estimation was performed in PASS (NCSS, LLC), targeting a two-sided 95% confidence interval half-width ≤0.10, indicating that approximately 300 anterior sites were required. Secondary binary outcomes was assessed using Wilson intervals. Assuming values of 0.70–0.85, approximately 500 sites were needed to achieve a half-width of ±7.5%, corresponding to a sample size of over 50 patients. Inclusion criteria & Exclusion criteria Participants were eligible if they met all of the following: Malocclusion type: Angle Class I or mild Class II/III malocclusion (≤1/4-cusp molar deviation). Malocclusion characteristic: Anterior dental crowding ≥4 mm in at least one arch. Treatment modality: Extraction of four first premolars followed by comprehensive fixed orthodontic treatment, with no major mid-course protocol modifications. Age and dentition: ≥18 years at treatment onset, full permanent dentition, and no active periodontal inflammation. Data completeness: High-quality digital intraoral scans (STL) available both pre-treatment (T0) and post-treatment (T1), with complete clinical records and photographs. Participants were excluded if they exhibited any of the following: Local surgical or restorative confounders: History of anterior periodontal or orthognathic surgery; traumatic injuries; esthetic restorations; dental implants; or tooth anomalies (missing/malformed/impacted incisors). Treatment-related modifications: Interproximal enamel reduction, skeletal anchorage devices, or removable appliances affecting the anterior segment. Periodontal impairment: Baseline alveolar bone or attachment loss in the anterior region, or clinically significant periodontal deterioration during treatment. Noncompliance or poor documentation: Treatment interruption, incomplete records, or inadequate scan quality preventing reliable papilla evaluation. MEASUREMENTS Actual postoperative OGEs were evaluated from standardized intraoral photographs obtained aftere treatment. OGE presence was recorded as a binary outcome (0=absent, 1=present). Severity was graded using a four-level scale modified from the Jemt papilla index, based on the vertical extent of interdental papilla fill (Fig. 2): score 0 (Jemt 3), complete papilla fill; score 1 (Jemt 2), >50% fill; score 2 (Jemt 1), <50% fill; and score 3 (Jemt 0), complete papilla loss. Predicted OGEs were evaluated on simulated final tooth positions generated from pre-treatment STL models using the iOrtho platform (Angelalign Technology Inc., Shanghai). Virtual setups were prepared by a certified technician, and binary and ordinal assessments were performed by an orthodontist using standardized criteria. Assessments at 10 anterior interproximal sites were conducted by a single calibrated examiner blinded to subgroup allocation, with 20% reassessed after 4 weeks to evaluate intra-examiner reliability. Fig.3 illustrates a representative clinical case comparing actual and predicted OGEs using intraoral photographs and digital setup simulation. STATISTICAL ANALYSIS All statistical analyses were performed in R (version 4.4.1; R Foundation for Statistical Computing, Vienna, Austria). Continuous variables are presented as mean ± SD and categorical variables as counts and percentages. Binary OGE prediction performance was assessed using accuracy, sensitivity, specificity, PPV, NPV, F1-score, and Cohen’s κ, overall and by arch, sex, and site. Ordinal agreement for Jemt grades was evaluated using weighted Cohen’s κ, mean absolute error (MAE), and Spearman’s correlation (ρ). Calibration was assessed using Jensen–Shannon divergence (JSD), with bias defined as the difference between predicted and observed OGE incidence. Statistical metrics and definitions are summarized in Table 1. 3. RESULTS 3.1 Participant characteristics A total of 62 adult patients undergoing fixed orthodontic treatment with four 1st premolar extractions were included in the analysis. The sample consisted of 52 females and 10 males. The mean age at treatment onset was 24.1 years. All participants met the predefined eligibility criteria, with complete digital records available for analysis. 3.2 Overall prediction agreement for OGE epidemiology and severity Across all anterior sites, the predicted incidence of open gingival embrasures (OGEs) showed close alignment with clinical observations (76.5% vs 71.9%; bias = + 4.5%), indicating no substantial systematic misestimation at the population level(Table 2 ). The severity-grade distribution demonstrated low Jensen–Shannon divergence (JSD = 0.004) and a strong ordinal correlation (ρ = 0.786), confirming good agreement in the directional trend of papilla loss. Exact-match accuracy reached 80.8%, with most discrepancies reflecting minor over-grading rather than under-grading (14.4% vs 4.8%). Subgroup comparison revealed slightly weaker calibration performance in the mandible compared with the maxilla (bias: +2.3% vs + 6.8%; JSD: 0.006 vs 0.004). Sex-based differences were minimal overall, though male patients demonstrated modestly larger over-prediction tendencies (21.0% vs 13.1%) despite similar rank correlations (ρ = 0.726 vs 0.800). Collectively, the digital setup provided a clinically acceptable representation of postoperative OGE prevalence and severity distribution, with prediction deviations largely within mild ranges across major subgroups. Table 2 Calibration and distribution agreement of OGEs prediction by subgroup. Group n Observed incidence Predicted incidence Bias JSD Spearman_rho Exact match Under Over Overall 620 71.9 76.5 4.5 0.004 0.786 80.8 4.8 14.4 Maxilla 310 57.7 64.5 6.8 0.004 0.740 82.6 4.5 12.9 Mandible 310 86.1 88.4 2.3 0.006 0.731 79.0 5.2 15.8 Female 520 73.5 76.7 3.3 0.003 0.800 81.9 5.0 13.1 Male 100 64.0 75.0 11.0 0.013 0.726 75.0 4.0 21.0 3.3 Binary prediction performance for OGE occurrence As shown in Table 3 , binary prediction demonstrated high overall performance, with an accuracy of 91.6%, balanced sensitivity and specificity (0.916 and 0.918), and a high PPV (97.3%) with lower NPV (77.0%). The F1-score was 0.943. Mandibular sites showed higher accuracy and sensitivity than maxillary sites, which exhibited slightly more missed detections, while sex-based differences were modest, with higher sensitivity in females and higher specificity in males. Table 3 Binary performance metrics for predicting the presence of OGEs. Group n Accuracy Sensitivity Specificity PPV NPV F1 Youden_J Overall 620 0.916 0.916 0.918 0.973 0.770 0.943 0.833 Maxilla 310 0.874 0.850 0.918 0.950 0.771 0.897 0.768 Mandible 310 0.958 0.964 0.917 0.989 0.767 0.976 0.880 Female 520 0.925 0.930 0.909 0.971 0.797 0.950 0.839 Male 100 0.870 0.840 0.960 0.984 0.667 0.906 0.800 Confusion matrices (Fig. 4 ), calibration plots (Fig. 5 ), and ROC curves (Fig. 6 ; AUCs > 0.8) confirmed stable and well-calibrated binary prediction across subgroups. 3.4 Agreement in OGE severity classification (ordinal Jemt grades) As shown in Table 4 , predicted and actual Jemt grades demonstrated moderate-to-substantial agreement (κ = 0.716–0.774) with a low MAE (0.194). Exact-grade agreement was achieved in 80.8% of sites, and 99.8% of discrepancies were within ± 1 grade. Table 4 Overall and subgroup agreement between predicted and actual OGE severity (Jemt grades 0–3). Group n Top-1 accuracy Top-1 ± 1 accuracy MAE Weighted Kappa (linear) Weighted Kappa (quadratic) Spearman rho Overall 620 0.808 0.998 0.194 0.716 0.774 0.786 Maxilla 310 0.826 0.997 0.177 0.689 0.716 0.740 Mandible 310 0.790 1.000 0.210 0.680 0.743 0.731 Female 520 0.819 1.000 0.181 0.733 0.790 0.800 Male 100 0.750 0.990 0.260 0.633 0.695 0.726 Agreement was slightly lower in mandibular sites and in male patients, although Spearman’s correlations remained strong across subgroups (ρ ≥ 0.726). Ordered confusion matrices (Fig. 7 ) showed predictions clustered along the diagonal, indicating predominantly minor misgrading. 3.5 Anatomical site–dependent prediction performance As shown in Tables 5 and 6 , prediction performance varied modestly by papilla site. Binary accuracy exceeded 85% at most anterior sites, with greater stability in maxillary central and canine-adjacent embrasures and more variability in mandibular sites. Severity misclassification was predominantly minor (± 1 grade). Table 5 Site-specific binary prediction performance for postoperative OGEs presence. Site type n Accuracy Sensitivity Specificity PPV NPV F1 Youden’s J 13 − 12 62 0.903 0.960 0.865 0.828 0.970 0.889 0.825 12 − 11 62 0.839 0.976 0.550 0.820 0.917 0.891 0.526 11–21 62 0.903 0.915 0.867 0.956 0.765 0.935 0.782 21–22 62 0.871 1.000 0.579 0.843 1.000 0.915 0.579 22–23 62 0.855 0.864 0.850 0.760 0.919 0.809 0.714 43 − 42 62 0.935 0.979 0.786 0.940 0.917 0.959 0.765 42 − 41 62 0.968 0.982 0.800 0.982 0.800 0.982 0.782 41 − 31 62 1.000 1.000 1.000 1.000 1.000 1.000 1.000 31–32 62 0.919 0.982 0.429 0.931 0.750 0.956 0.410 32–33 62 0.968 1.000 0.833 0.962 1.000 0.980 0.833 Table 6 Site-specific agreement between predicted and actual OGEs severity (Jemt grades 0–3). Site type n Top-1 accuracy Top-1 ±1 accuracy MAE Weighted Kappa (linear) Weighted Kappa (quadratic) Spearman rho 13 − 12 62 0.903 1.000 0.097 0.816 0.828 0.819 12 − 11 62 0.694 1.000 0.306 0.472 0.559 0.604 11–21 62 0.871 1.000 0.129 0.744 0.774 0.773 21–22 62 0.839 1.000 0.161 0.631 0.656 0.696 22–23 62 0.823 0.984 0.194 0.628 0.613 0.679 43 − 42 62 0.839 1.000 0.161 0.684 0.729 0.761 42 − 41 62 0.694 1.000 0.306 0.527 0.610 0.576 41 − 31 62 0.726 1.000 0.274 0.580 0.667 0.580 31–32 62 0.758 1.000 0.242 0.625 0.694 0.697 32–33 62 0.935 1.000 0.065 0.854 0.870 0.877 Site-specific confusion matrices showed stronger agreement in maxillary anterior regions (Fig. 4 ), whereas mandibular central–lateral sites exhibited greater dispersion (Fig. 7 ). Overall, digital setup–based site-specific prediction was clinically reliable, while highlighting mandibular incisor regions as requiring closer esthetic monitoring. 4. DISCUSSION This retrospective study evaluated whether digital orthodontic setups, originally designed for clear aligner planning, can serve as a predictive tool for postoperative open gingival embrasure (OGE) development in adult extraction cases treated with fixed appliances. The results demonstrated that digital setups provided clinically acceptable prediction accuracy for both the presence and severity of papilla deficiency. Binary performance remained high across subgroups, with strong sensitivity indicating reliable detection of sites at risk. Severity estimates based on the modified Jemt index achieved substantial ordinal agreement, with nearly all discrepancies limited to minor ± 1–grade differences. Site-based patterns highlighted the mandibular anterior region as the area with the greatest variability, consistent with its biomechanical vulnerability. Collectively, these findings support the potential of digital setups to assist clinicians in anticipating soft-tissue esthetic risks during treatment planning. The predictive value of digital orthodontic setups for OGE development aligns with established morphologic determinants of papilla loss. Prior research demonstrates that an ICP–AC distance exceeding 5 mm, small increases in vertical discrepancies (ΔICP–AC, ΔCEJ–AC), and movement thresholds such as excessive intrusion or labial inclination significantly elevate OGE risk[ 2 ]. Because these geometric relationships—contact-point position, crown form, root alignment, and interproximal spacing—are inherently embedded within digital setups, the simulated final dentition can approximate the anatomical configurations most susceptible to papillary insufficiency, even without explicit soft-tissue modeling. From a biological point, adults undergoing orthodontic treatment are inherently more susceptible to papillary recession because of age-related alveolar remodeling and diminished regenerative capacity, as papilla height decreased 0.012 mm each year of increasing age according to a published research[ 10 ]. Extraction therapy and crowding further exacerbate this vulnerability by requiring substantial tooth movement, root realignment, and space closure. As shown in previous evidences, the prevalence of OGEs is relatively high and strongly associated with tooth extraction, with severity decreasing in the order of one-lower-incisor extraction > two-premolar extraction > non-extraction cases[ 2 ]. Meanwhile, greater mandibular crowding was identified as an independent risk factor for OGE occurrence (OR = 0.846) in particular[ 12 ]. The anterior region, especially the mandibular incisors, is anatomically predisposed to OGE due to thin labial bone plates, limited keratinized tissue, and high sensitivity to even minor changes in root positioning[ 3 , 13 ]. These biological tendencies are fully reflected in our study population—adult extraction cases with ≥ 4 mm crowding—and correspond with the high overall OGE prevalence observed. Moreover, the site-dependent differences detected in our measurements, with mandibular anterior sites showing greater susceptibility, reinforce the established notion that local anatomy strongly influences papillary stability. Several clinical studies have suggested that clear aligner therapy is associated with a higher risk of open gingival embrasures than fixed appliances. Yang et al. reported that the incidence of OGEs between maxillary and mandibular central incisors was 35.0% and 38.0% in the clear-aligner group, compared with 18.0% and 24.0% in the fixed-appliance group[ 21 ]. Cui et al. further showed that in non-extraction Invisalign cases the overall anterior OGE incidence reached 13.4% in the maxilla and 30.7% in the mandible, with the highest rate (≈ 39%) between mandibular central incisors, and identified age, mandibular crowding and the distance from the interproximal contact point to the alveolar crest as independent risk factors[ 12 ]. These findings are consistent with earlier work on fixed appliances[ 3 , 22 ], which emphasized the roles of anterior crowding, root divergence and contact-point–bone relationships, but they also highlight that the “magnitude” and “frequency” of OGEs appear greater in clear-aligner cohorts. Our results mirror this pattern at the level of digital prediction. Across the whole sample, the digitally generated setups systematically predicted a higher incidence and greater severity of OGEs than were actually observed after fixed-appliance treatment, both in overall analyses and within arch- and sex-stratified subgroups. Site-specific comparisons showed that predicted OGEs were particularly over-represented in mandibular anterior sites—especially between the central incisors and adjacent contacts—where clinical OGEs did occur but at a lower frequency. In terms of error structure, false-positive predictions clearly outnumbered false negatives in most subgroups, indicating that the digital setup behaves as a conservative “over-screening” tool for black triangles rather than an exact replica of the final soft-tissue outcome. This tendency is fully coherent with the aligner literature, in which mandibular anterior embrasures are consistently identified as the most vulnerable sites. The overestimation pattern may be partially explained by how aligner-based digital setups and algorithms encode risk. As discussed by previous studies[ 4 , 12 , 21 ], clear aligner systems often involve extensive labial and vertical movements, multiple attachments and prolonged coverage of the gingival margin; the aligner material extends apical to the contact point and may occupy the interdental space during papilla remodeling. These appliance-specific features, together with high ICP–ABC distances and triangular crown forms, are implicitly built into the virtual setup and may bias the learned prediction models toward expecting more papillary deficiency than actually occurs under fixed-appliance biomechanics. In our study, however, all patients were treated with conventional fixed appliances, which may allow slightly better papilla preservation in some sites, leading to a gap between “virtual” and real soft-tissue outcomes. This bias towards overprediction has two important clinical implications. First, it suggests that setup-based OGE estimates should be understood as “upper-bound risk scenarios”: a high predicted probability or grade of OGE flags sites that truly deserve attention, but a proportion of these will not develop clinically visible black triangles. Second, the discrepancy between predicted and observed outcomes underscores the need to further refine current algorithms—ideally by training on large, appliance-specific datasets and by explicitly incorporating known risk modifiers such as extraction pattern, crowding resolution strategy and periodontal phenotype. Until then, clinicians should interpret digital OGE predictions qualitatively, integrating them with clinical judgement. Several limitations should be acknowledged when interpreting the present findings. First, this study adopted a retrospective design with data collected from a single clinical center, which may limit generalizability across populations with different periodontal phenotypes or treatment protocols. The sample primarily included adult extraction cases with moderate crowding; therefore, outcomes may differ in non-extraction or adolescent cohorts where papillary resilience and bone morphology vary. Second, although the predictive setup was generated by trained technicians using standardized digital protocols, the virtual alignment still lacks direct modeling of soft-tissue elasticity, gingival thickness, and alveolar crest remodeling—factors that could influence papilla fill in vivo. Third, the Jemt index, while widely used and reproducible, remains a semi-quantitative visual scale and may not fully capture three-dimensional papilla morphology or subtle contour discrepancies[ 23 ]. Integration with volumetric soft-tissue analysis or digital photogrammetry may enhance assessment precision in future research. Traditional methods for predicting the risk of open gingival embrasures rely on extensive clinical data collection—patient age, crown morphology, alveolar and gingival biotype, root length and parallelism, tooth-movement patterns, periodontal probing measurements, and geometric calculations of contact-point–crest relationships. Although informative, these approaches are analytically complex, time-consuming, and often unintuitive for both clinicians and patients. In contrast, digital orthodontic setups offer a data-driven alternative: machine-learning and deep-learning algorithms can synthesize large amounts of geometric and morphological information and output visually interpretable predictions directly from the “black box.” A scoping review reported that AI algorithms already achieve near-expert performance in tooth segmentation, CBCT–IOS registration, and digital setup prediction, and can provide automated, visually interpretable outputs that support treatment planning and remote monitoring[ 24 ]. Similar conclusions were drawn in several reviews[ 25 , 26 ], which highlight the potential of AI to enhance efficiency and predictive accuracy in orthodontics. Against this backdrop, our study can be viewed as a posterior validation of the soft-tissue predictions generated by a commercial clear-aligner platform, demonstrating that digitally derived papilla simulations and OGE severity estimates align reasonably well with real clinical outcomes. These findings underscore the potential of AI-enhanced digital setups in soft-tissue outcome prediction. Future research should incorporate a more comprehensive set of established OGE risk factors to build advanced machine-learning models, and aligner companies may further refine their digital setup interfaces to improve the accuracy of soft-tissue rendering and prognostic reliability. Overall, while the current approach demonstrates promising feasibility for soft-tissue risk visualization in fixed orthodontics, its translation to routine clinical practice requires further validation, model refinement, and standardization across different digital platforms. 5. CONCLUSION Digital orthodontic setups derived from pre-treatment intraoral scans offer a feasible adjunct for predicting postoperative OGEs in adult extraction patients treated with fixed appliances, supporting their role in soft-tissue esthetic risk assessment and treatment planning. Declarations All authors contributed to the study conception and design. All authors read and approved the final manuscript. AUTHOR CONTRIBUTIONS Xinyi, Ren: Methodology, Investigation, Visualization, Writing – original draft. Jiaxin, Deng: Data curation, Formal analysis, Software. Lang Lei: Conceptualization, Writing – review & editing. Jialing, Li: Writing – review & editing, Validation. Huang Li: Supervision, Project administration. ACKNOWLEDGMENTS This research was supported by Nanjing Health Development Key Project (ZKX23055). The authors thank the orthodontic technicians for their assistance with digital setup preparation. No external professional writing, editing, or proofreading services were used for this manuscript. Ethics approval This retrospective study was approved by the Ethics Committee of Nanjing Stomatological Hospital, Nanjing University (Approval No. NJSH-2023NL-036). Consent to participate The requirement for informed consent was waived by the Ethics Committee of Nanjing Stomatological Hospital, Nanjing University (Approval No. NJSH-2023NL-036) due to the retrospective design of the study. Consent to publish Not applicable. Conflict of Interest The authors declare that they have no conflict of interest. Funding This work was supported by Nanjing Health Development Key Project (ZKX23055). Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Author contributions Xinyi, Ren: Methodology, Investigation, Visualization, Writing – original draft. Jiaxin, Deng: Data curation, Formal analysis, Software. Lang Lei: Conceptualization, Writing – review & editing. Jialing, Li: Writing – review & editing, Validation. Huang Li: Supervision, Project administration. References J R, K. & V G, K. Open gingival embrasures after orthodontic treatment in adults: prevalence and etiology. Am J Orthod Dentofacial Orthop 120 (2001). https://doi.org/10.1067/mod.2001.114831 Zhang, Y. et al. Clear aligners and open gingival embrasures: Retrospective study of epidemiology and risk factors. J Periodontol (2025). https://doi.org/10.1002/jper.11373 An, S. S., Choi, Y. J., Kim, J. Y., Chung, C. J. & Kim, K. H. Risk factors associated with open gingival embrasures after orthodontic treatment. Angle Orthod 88, 267–274 (2018). https://doi.org/10.2319/061917-399.12 Zhang, Y. et al. IPR treatment and attachments design in clear aligner therapy and risk of open gingival embrasures in adults. Prog Orthod 24, 1 (2023). https://doi.org/10.1186/s40510-022-00452-1 Erkang, T. et al. Factors influencing open gingival embrasures in orthodontic treatment: a retrospective clinical study. Prog Orthod 26 (2025). https://doi.org/10.1186/s40510-025-00554-6 Sanabel O, B. Interdental papilla recession and reconstruction of the lost triangle: a review of the current literature. Front Dent Med 5 (2025). https://doi.org/10.3389/fdmed.2024.1537452 Mahmoud K, A.-O. et al. The knowledge regarding the impacts and management of black triangles among dental professionals and laypeople. Sci Rep 14 (2024). https://doi.org/10.1038/s41598-024-61356-0 Yudai, O. et al. Interdisciplinary Management of Open Interproximal Embrasures. J Esthet Restor Dent (2025). https://doi.org/10.1111/jerd.70015 Yubohan, Z. et al. Biomechanical factors in the open gingival embrasure region during the intrusion of mandibular incisors: A new model through finite element analysis. Front Bioeng Biotechnol 11 (2023). https://doi.org/10.3389/fbioe.2023.1149472 Chow, Y. C., Eber, R. M., Tsao, Y. P., Shotwell, J. L. & Wang, H. L. Factors associated with the appearance of gingival papillae. J Clin Periodontol 37, 719–727 (2010). https://doi.org/10.1111/j.1600-051X.2010.01594.x Jung, J. S., Lim, H. K., Lee, Y. S. & Jung, S. K. The Occurrence and Risk Factors of Black Triangles Between Central Incisors After Orthodontic Treatment. Diagnostics (Basel) 14 (2024). https://doi.org/10.3390/diagnostics14232747 Cui, W., Liu, Y., Zhao, Y., Lei, L. & Li, H. Risk factors for open gingival embrasures after clear aligners treatment: a retrospective study. BMC Oral Health 25, 547 (2025). https://doi.org/10.1186/s12903-025-05915-5 Kim, T., Miyamoto, T., Nunn, M. E., Garcia, R. I. & Dietrich, T. Root proximity as a risk factor for progression of alveolar bone loss: the Veterans Affairs Dental Longitudinal Study. J Periodontol 79, 654–659 (2008). https://doi.org/10.1902/jop.2008.070477 Mahasneh, S. A., Goodwin, M., Pretty, I. & Cunliffe, J. The use of radiographs to assess the impact of the distance between the contact area and the crest of the bone to predict the presence or absence of interdental papilla: an in vivo study. Br Dent J (2023). https://doi.org/10.1038/s41415-023-6184-z Guo, F. et al. Invisalign ClinCheck can predict open gingival embrasures in adult extraction cases: a pilot study. Angle Orthod 95, 389–396 (2025). https://doi.org/10.2319/091424-752.1 Lindsay, R. et al. Effectiveness of clear aligner therapy for orthodontic treatment: A systematic review. Orthod Craniofac Res 23 (2019). https://doi.org/10.1111/ocr.12353 Linghuan, R. et al. The predictability of orthodontic tooth movements through clear aligner among first-premolar extraction patients: a multivariate analysis. Prog Orthod 23 (2022). https://doi.org/10.1186/s40510-022-00447-y Gabriele, R., Simone, P., Tommaso, C., Andrea, D. & Cesare L, D. Efficacy of clear aligners in controlling orthodontic tooth movement: a systematic review. Angle Orthod 85 (2014). https://doi.org/10.2319/061614-436.1 Tommaso, C. et al. Predictability of orthodontic tooth movement with aligners: effect of treatment design. Prog Orthod 24 (2023). https://doi.org/10.1186/s40510-022-00453-0 Abdalrahman Mohieddin, K., Kinda, S., Mohammad Younis, H. & Nikolaos, G. Digital setup accuracy for moderate crowding correction with fixed orthodontic appliances: a prospective study. Prog Orthod 25 (2024). https://doi.org/10.1186/s40510-024-00513-7 Tianrui, Y. et al. The incidence and severity of open gingival embrasures in adults treated with clear aligners and fixed appliances: a retrospective cohort study. Head Face Med 19 (2023). https://doi.org/10.1186/s13005-023-00375-0 Tadataka, I., Masaru, Y., Daijiro, M. & Kazutaka, K. Prediction and causes of open gingival embrasure spaces between the mandibular central incisors following orthodontic treatment. Aust Orthod J 20 (2006). Daniele, C., Stefania, R. & Giuseppe, C. The Papilla Presence Index (PPI): a new system to assess interproximal papillary levels. Int J Periodontics Restorative Dent 24 (2004). https://doi.org/10.11607/prd.00.0596 Débora Costa, R. et al. Unveiling the role of artificial intelligence applied to clear aligner therapy: A scoping review. J Dent 154 (2025). https://doi.org/10.1016/j.jdent.2025.105564 Sanjeev B, K. et al. Developments, application, and performance of artificial intelligence in dentistry - A systematic review. J Dent Sci 16 (2021). https://doi.org/10.1016/j.jds.2020.06.019 Andrej, T., Veronika, K. & Ivan, V. Artificial Intelligence in Orthodontic Smart Application for Treatment Coaching and Its Impact on Clinical Performance of Patients Monitored with AI-TeleHealth System. Healthcare (Basel) 9 (2021). https://doi.org/10.3390/healthcare9121695 Table Table 1 is available in the Supplementary Files section Additional Declarations No competing interests reported. Supplementary Files Table1.docx 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-8439148","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":572830915,"identity":"cffdd7b5-e707-4bcd-9310-51e747b2a203","order_by":0,"name":"Xinyi Ren","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Research Institute of Stomatology, Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Xinyi","middleName":"","lastName":"Ren","suffix":""},{"id":572830916,"identity":"da355836-03ef-4346-8a44-db6519b5877b","order_by":1,"name":"Jiaxin Deng","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Research Institute of Stomatology, Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Jiaxin","middleName":"","lastName":"Deng","suffix":""},{"id":572830917,"identity":"3133fb6e-5d8f-460c-9795-0726950be4d2","order_by":2,"name":"Lang Lei","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Research Institute of Stomatology, Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Lang","middleName":"","lastName":"Lei","suffix":""},{"id":572830920,"identity":"f6689ca7-0689-480e-ad6a-50330fb5df9c","order_by":3,"name":"Jialing Li","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Research Institute of Stomatology, Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Jialing","middleName":"","lastName":"Li","suffix":""},{"id":572830924,"identity":"76dd4c14-829e-4af9-ac94-2b004988632c","order_by":4,"name":"Huang Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4ElEQVRIiWNgGAWjYBACA2YeMF3fz97AYABkMDYQq4VxZs8BYrUwQLVsmJEAYRDWws57TOLnjlpmA8m3B4p5GGxkNxxgfvYAv8P40iR7zxxnM5fOSzDmYUgz3nCAzdyAgF/MJHjbjvFYzs4xAGo5nLjhAA+bBCEtkn/bjkkY3DwD0vKfOC3SvG01BgY3eEBaDhClxdhatu1AgmRPjoHhHINk45mH2czwarHvP2N4821bXQI/+xkzgzcVdrJ9x5uf4dUCBYdBBJsBODKZiVAPBHUggvkBcYpHwSgYBaNgpAEAMiJAw0ycxcUAAAAASUVORK5CYII=","orcid":"","institution":"Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Research Institute of Stomatology, Nanjing University","correspondingAuthor":true,"prefix":"","firstName":"Huang","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2025-12-24 05:38:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8439148/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8439148/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100070219,"identity":"e643a4ff-7e45-4d60-aacc-8049ad2ca5c2","added_by":"auto","created_at":"2026-01-12 16:17:06","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1671262,"visible":true,"origin":"","legend":"","description":"","filename":"COImanuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/07bbbde994fe458678dea58c.docx"},{"id":100070229,"identity":"016e53e6-d015-404a-8c0b-923f8bd716c6","added_by":"auto","created_at":"2026-01-12 16:17:08","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6775,"visible":true,"origin":"","legend":"","description":"","filename":"4bddf32368be49bdaa0549dd3baf3ccd.json","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/dcb59f70fa682dc8bca7fdda.json"},{"id":100070207,"identity":"936e2075-bc88-42f3-b38b-c80ce6e7f7e2","added_by":"auto","created_at":"2026-01-12 16:17:03","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":124518,"visible":true,"origin":"","legend":"","description":"","filename":"4bddf32368be49bdaa0549dd3baf3ccd1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/dd1d8b5c639a5f746af9fca6.xml"},{"id":100070204,"identity":"8ca59713-37ad-4817-bfa6-3aaf780889c0","added_by":"auto","created_at":"2026-01-12 16:17:02","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128297,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/9c9836e9cf70308b42b8f08b.png"},{"id":100070205,"identity":"ab9bc154-d686-45d8-adb5-e87416c18adf","added_by":"auto","created_at":"2026-01-12 16:17:02","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":372076,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/3a7e61bd26affbc24d0decc2.png"},{"id":100070245,"identity":"86f2817c-dc77-4da5-977d-45593dbe331b","added_by":"auto","created_at":"2026-01-12 16:17:16","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":624066,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/dff3d77e98ce153810123df6.png"},{"id":100070246,"identity":"3f7684ae-b324-42ce-be9d-98b4e589f408","added_by":"auto","created_at":"2026-01-12 16:17:16","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":112418,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/99b813fe13f4a6fa487adc74.png"},{"id":100070227,"identity":"62074238-d599-45b6-958e-e3352645d6e3","added_by":"auto","created_at":"2026-01-12 16:17:07","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":87632,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/e3b4009c79adc76b42219944.png"},{"id":100070220,"identity":"e4ae5bc9-3339-45ce-b69f-7b5362964ce7","added_by":"auto","created_at":"2026-01-12 16:17:06","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":88496,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/2b2ea5f2752fd095b8e9413f.png"},{"id":100070196,"identity":"0f07d43d-86bc-424c-9bd3-e5347f8bedab","added_by":"auto","created_at":"2026-01-12 16:16:55","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":268310,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/9e4e2881a35ed37db65dfadb.png"},{"id":100070224,"identity":"83d282bf-e227-43f4-ac35-978af91471f6","added_by":"auto","created_at":"2026-01-12 16:17:07","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18202,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/66ff59a03428235e5e02f1d6.png"},{"id":100070203,"identity":"95240313-1686-4c9a-9ab5-fff05bfc8514","added_by":"auto","created_at":"2026-01-12 16:17:02","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64706,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/828623f52c8e061f725db56d.png"},{"id":100070198,"identity":"3421342d-b0f1-41de-a731-259f4c731286","added_by":"auto","created_at":"2026-01-12 16:16:55","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":81634,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/2878f11a63be8a3814b3952f.png"},{"id":100070239,"identity":"602e530b-ab2d-4997-9658-edfba96ace32","added_by":"auto","created_at":"2026-01-12 16:17:10","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":34195,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/e8d20517416d77d0dd9e5338.png"},{"id":100070213,"identity":"d130a855-ac3b-4469-8be8-4496a37538e3","added_by":"auto","created_at":"2026-01-12 16:17:04","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21912,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/c5a5e95cddba431fb325e703.png"},{"id":100070217,"identity":"5f6c2a2f-be12-4ccc-99ef-abe8646a53e1","added_by":"auto","created_at":"2026-01-12 16:17:06","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":41279,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/e08ebee15b4aa03a3e734200.png"},{"id":100070208,"identity":"1b38ad78-84f6-4d0c-a427-4359a6322526","added_by":"auto","created_at":"2026-01-12 16:17:04","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":124164,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/df143678d404eeb1447d492c.png"},{"id":100070218,"identity":"db3222d3-c6b1-48c5-830a-0890fe510bea","added_by":"auto","created_at":"2026-01-12 16:17:06","extension":"xml","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":119836,"visible":true,"origin":"","legend":"","description":"","filename":"4bddf32368be49bdaa0549dd3baf3ccd1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/b7edc119a966f3991cd4c79f.xml"},{"id":100364381,"identity":"2679a17f-b616-4efb-9304-85fe9d47e0f4","added_by":"auto","created_at":"2026-01-16 07:53:35","extension":"html","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133521,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/2bc0b7f7a395980b92471b06.html"},{"id":100070215,"identity":"c10a16c2-eae8-4c5f-944a-b946c1800b0f","added_by":"auto","created_at":"2026-01-12 16:17:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":159735,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of sample screening and enrollment\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/60417f7345fb564f7147027b.png"},{"id":100070210,"identity":"4dd34f43-66fe-4ab0-bae2-0d51111a1d96","added_by":"auto","created_at":"2026-01-12 16:17:04","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":245508,"visible":true,"origin":"","legend":"\u003cp\u003eSeverity Grading of OGEs (Based on Jemt Index). Line A was drawn through the contact point of adjacent teeth. Line C passed through the highest buccal gingival curvature of the crowns. Line B, drawn parallel to Line A, bisected the perpendicular distance between Lines A and C, representing half of the gingival papilla height\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/c9ddd872d7022675411ecbd2.png"},{"id":100070195,"identity":"0f419b5f-515b-4142-a26e-46e56a616cd5","added_by":"auto","created_at":"2026-01-12 16:16:55","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":413056,"visible":true,"origin":"","legend":"\u003cp\u003eA representative case of actual and digitally predicted OGEs. (a) Pre-treatment photograph. (b) Post-treatment photograph. (c) Pre-treatment digital model. (d) Simulated final tooth position for OGE assessment\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/4c5b9dea3b11a2d645263f85.png"},{"id":100365279,"identity":"3af203ee-a19b-444f-a4b0-60b5090518e2","added_by":"auto","created_at":"2026-01-16 07:54:57","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":106267,"visible":true,"origin":"","legend":"\u003cp\u003eConfusion-matrix heatmaps for binary prediction of postoperative OGEs (1 = present, 0 = absent). Cell values represent site counts for each prediction–observation combination, with darker shading along the diagonal indicating precise predictions\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/724573b20303fe01912f6a08.png"},{"id":100070192,"identity":"6afa8bcb-506b-43a3-bb7b-97627adb0833","added_by":"auto","created_at":"2026-01-12 16:16:54","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":68067,"visible":true,"origin":"","legend":"\u003cp\u003eCalibration plots for binary prediction of postoperative OGEs presence. (a) Overall. (b) By arch (maxilla vs mandible). (c) By sex (female vs male). Points indicate observed OGE incidence for each predicted category (0 = absent, 1 = present) with 95% Wilson confidence intervals; the dashed diagonal line represents perfect calibration\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/845fca6d3792762ab5c6e38c.png"},{"id":100070231,"identity":"a7fa491a-fa4f-40e4-837c-a77d80b0343a","added_by":"auto","created_at":"2026-01-12 16:17:08","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":83732,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristic (ROC) curves for binary prediction of postoperative OGEs prediction. (a) Overall. (b) By arch (maxilla vs mandible). () By sex (female vs male). The diagonal dashed line indicates no discrimination, and the area under the curve (AUC) shown in each panel quantifies classification performance\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/47262d9e3b3f9b87aa86b43d.png"},{"id":100070221,"identity":"894cf01d-c48a-4e84-8233-47e72a0fffd6","added_by":"auto","created_at":"2026-01-12 16:17:07","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":116496,"visible":true,"origin":"","legend":"\u003cp\u003eOrdered confusion matrices comparing predicted and actual postoperative Jemt grades (0–3). (a) Overall matrix. (b) By dental arch (maxilla vs mandible). (c) By sex (female vs male). (D) Site-specific matrices for 10 anterior interproximal sites. Cell values indicate case counts, with darker shading representing higher frequency. The diagonal denotes perfect agreement; off-diagonal cells indicate over- or under-grading\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/52c49da5147dc6dfcbfaf959.png"},{"id":104403580,"identity":"ec7f758c-7fcd-44b0-9f77-cce6a006053e","added_by":"auto","created_at":"2026-03-11 12:18:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2277885,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/175c6293-42b7-4e37-9fb2-280ea2798804.pdf"},{"id":100070216,"identity":"2d315e9b-993e-40b1-9560-197fc222651b","added_by":"auto","created_at":"2026-01-12 16:17:05","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":17352,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.docx","url":"https://assets-eu.researchsquare.com/files/rs-8439148/v1/ed837d443c3f6e187b43c9e8.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Digital Orthodontic Setup as a Predictive Tool for Post-treatment Open Gingival Embrasures in Adult Extraction Cases with Fixed Appliances: A Retrospective Study","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003eOpen gingival embrasures (OGEs), commonly referred to as \u0026ldquo;black triangles\u0026rdquo;, are a frequent esthetic and functional concern following orthodontic treatment, particularly in adult patients[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Epidemiological studies have reported OGE prevalence ranging from approximately 20% to 38%, with higher occurrence in the anterior region with triangular tooth morphology, dental crowding, or compromised periodontal support[\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Characterized by incomplete interdental papilla fill, OGEs negatively affect smile esthetics and may promote plaque accumulation, food impaction, and phonetic difficulties, thereby reducing periodontal stability and patient satisfaction[\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. As esthetic expectations among adult orthodontic patients continue to rise, even minor papillary deficiencies may disproportionately influence perceived treatment success[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Extraction therapy further increases susceptibility to OGEs because it involves greater tooth movement and alveolar bone remodeling[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Therefore, prediction of OGEs during treatment planning is essential to reduce esthetic risk, particularly in adult extraction cases.\u003c/p\u003e \u003cp\u003eThe development of OGEs is multifactorial and closely related to anatomical, periodontal, and biomechanical determinants[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Widely recognized risk factors include patient age, gingival biotype, interdental bone height, crown morphology, root angulation, and the vertical distance between the contact point and the alveolar crest[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan additionalcitationids=\"CR11 CR12\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Post-space-closure root divergence and reduced interdental contact area have also been associated with papillary loss[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Despite these established predictors, clinical risk assessment remains largely qualitative and relies on static two-dimensional records such as intraoral photographs and periapical radiographs[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. These methods provide limited information on three-dimensional interdental geometry and cannot simulate tooth movement-related changes, leading clinicians to depend heavily on experience-based judgment.\u003c/p\u003e \u003cp\u003eDigital orthodontic setups, originally developed for clear aligner therapy, enable accurate three-dimensional reconstruction and virtual simulation of tooth movement. Previous studies have demonstrated acceptable predictability of aligner-based setups for various tooth movements, including arch expansion, molar distalization, alignment efficiency, torque control, and intrusion[\u003cspan additionalcitationids=\"CR17 CR18\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. However, existing research has largely focused on hard-tissue, with limited attention to soft-tissue esthetics such as interdental papilla, and setup-based prediction in fixed appliance therapy remains underexplored. A recent prospective study reported acceptable average agreement but substantial individual discrepancies between planned and achieved tooth positions in fixed appliance treatment[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. In addition, our pilot study showed that Invisalign ClinCheck could reasonably predict postoperative OGEs treated with clear aligners[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Whether such predictive validity extends to fixed orthodontic treatment has not been characterized.\u003c/p\u003e \u003cp\u003eGeometric determinants encoded within digital setups\u0026mdash;such as contact-point position, crown morphology, and root alignment\u0026mdash;are known to be closely associated with papilla presence[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Therefore, digital setups may function for estimating papillary risk, even in the absence of explicit soft-tissue simulation. Notably, our prior study showed that clear aligners were associated with a higher incidence and severity of OGEs than fixed appliances[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], underscoring the importance of appliance-specific validation.\u003c/p\u003e \u003cp\u003eTherefore, this retrospective study evaluated the predictive validity of digital setups for postoperative OGEs in adult extraction patients treated with fixed appliances, using binary and ordinal (Jemt) outcomes with subgroup analyse. This study seeks to support the use of digital setup-based assessment in orthodontic soft-tissue esthetic risk evaluation.\u003c/p\u003e"},{"header":"2. MATERIALS AND METHODS","content":"\u003cp\u003ePARTICIPANTS\u003c/p\u003e\n\u003cp\u003ePatients\u003c/p\u003e\n\u003cp\u003eThis retrospective study was approved by the Ethics Committee of Nanjing Stomatological Hospital, Nanjing University (Approval No. NJSH-2023NL-036). Adult patients who initiated fixed orthodontic treatment between January 2020 and December 2022 at the same institution were consecutively screened (Fig.1).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSample size and power\u003c/p\u003e\n\u003cp\u003eA precision-based design was used, with overall weighted Cohen\u0026rsquo;s kappa (\u0026kappa;w) for Jemt grades (0-3) as the primary endpoint. Sample size estimation was performed in PASS (NCSS, LLC), targeting a two-sided 95% confidence interval half-width \u0026le;0.10, indicating that approximately 300 anterior sites were required. Secondary binary outcomes was assessed using Wilson intervals. Assuming values of 0.70\u0026ndash;0.85, approximately 500 sites were needed to achieve a half-width of \u0026plusmn;7.5%, corresponding to a sample size of over 50 patients.\u003c/p\u003e\n\u003cp\u003eInclusion criteria \u0026amp; Exclusion criteria\u003c/p\u003e\n\u003cp\u003eParticipants were eligible if they met all of the following:\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003eMalocclusion type: Angle Class I or mild Class II/III malocclusion (\u0026le;1/4-cusp molar deviation).\u003c/li\u003e\n \u003cli\u003eMalocclusion characteristic: Anterior dental crowding \u0026ge;4 mm in at least one arch.\u003c/li\u003e\n \u003cli\u003eTreatment modality: Extraction of four first premolars followed by comprehensive fixed orthodontic treatment, with no major mid-course protocol modifications.\u003c/li\u003e\n \u003cli\u003eAge and dentition: \u0026ge;18 years at treatment onset, full permanent dentition, and no active periodontal inflammation.\u003c/li\u003e\n \u003cli\u003eData completeness: High-quality digital intraoral scans (STL) available both pre-treatment (T0) and post-treatment (T1), with complete clinical records and photographs.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eParticipants were excluded if they exhibited any of the following:\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003eLocal surgical or restorative confounders: History of anterior periodontal or orthognathic surgery; traumatic injuries; esthetic restorations; dental implants; or tooth anomalies (missing/malformed/impacted incisors).\u003c/li\u003e\n \u003cli\u003eTreatment-related modifications: Interproximal enamel reduction, skeletal anchorage devices, or removable appliances affecting the anterior segment.\u003c/li\u003e\n \u003cli\u003ePeriodontal impairment: Baseline alveolar bone or attachment loss in the anterior region, or clinically significant periodontal deterioration during treatment.\u003c/li\u003e\n \u003cli\u003eNoncompliance or poor documentation: Treatment interruption, incomplete records, or inadequate scan quality preventing reliable papilla evaluation.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eMEASUREMENTS\u003c/p\u003e\n\u003cp\u003eActual postoperative OGEs were evaluated from standardized intraoral photographs obtained aftere treatment. OGE presence was recorded as a binary outcome (0=absent, 1=present). Severity was graded using a four-level scale modified from the Jemt papilla index, based on the vertical extent of interdental papilla fill (Fig. 2): score 0 (Jemt 3), complete papilla fill; score 1 (Jemt 2), \u0026gt;50% fill; score 2 (Jemt 1), \u0026lt;50% fill; and score 3 (Jemt 0), complete papilla loss.\u003c/p\u003e\n\u003cp\u003ePredicted OGEs were evaluated on simulated final tooth positions generated from pre-treatment STL models using the iOrtho platform (Angelalign Technology Inc., Shanghai). Virtual setups were prepared by a certified technician, and binary and ordinal assessments were performed by an orthodontist using standardized criteria. Assessments at 10 anterior interproximal sites were conducted by a single calibrated examiner blinded to subgroup allocation, with 20% reassessed after 4 weeks to evaluate intra-examiner reliability. Fig.3 illustrates a representative clinical case comparing actual and predicted OGEs using intraoral photographs and digital setup simulation.\u003c/p\u003e\n\u003cp\u003eSTATISTICAL ANALYSIS\u003c/p\u003e\n\u003cp\u003eAll statistical analyses were performed in R (version 4.4.1; R Foundation for Statistical Computing, Vienna, Austria). Continuous variables are presented as mean \u0026plusmn; SD and categorical variables as counts and percentages. Binary OGE prediction performance was assessed using accuracy, sensitivity, specificity, PPV, NPV, F1-score, and Cohen\u0026rsquo;s \u0026kappa;, overall and by arch, sex, and site. Ordinal agreement for Jemt grades was evaluated using weighted Cohen\u0026rsquo;s \u0026kappa;, mean absolute error (MAE), and Spearman\u0026rsquo;s correlation (\u0026rho;). Calibration was assessed using Jensen\u0026ndash;Shannon divergence (JSD), with bias defined as the difference between predicted and observed OGE incidence. Statistical metrics and definitions are summarized in Table 1.\u003c/p\u003e"},{"header":"3. RESULTS","content":"\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Participant characteristics\u003c/h2\u003e \u003cp\u003eA total of 62 adult patients undergoing fixed orthodontic treatment with four 1st premolar extractions were included in the analysis. The sample consisted of 52 females and 10 males. The mean age at treatment onset was 24.1 years. All participants met the predefined eligibility criteria, with complete digital records available for analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Overall prediction agreement for OGE epidemiology and severity\u003c/h2\u003e \u003cp\u003eAcross all anterior sites, the predicted incidence of open gingival embrasures (OGEs) showed close alignment with clinical observations (76.5% vs 71.9%; bias\u0026thinsp;=\u0026thinsp;+\u0026thinsp;4.5%), indicating no substantial systematic misestimation at the population level(Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The severity-grade distribution demonstrated low Jensen\u0026ndash;Shannon divergence (JSD\u0026thinsp;=\u0026thinsp;0.004) and a strong ordinal correlation (ρ\u0026thinsp;=\u0026thinsp;0.786), confirming good agreement in the directional trend of papilla loss. Exact-match accuracy reached 80.8%, with most discrepancies reflecting minor over-grading rather than under-grading (14.4% vs 4.8%).\u003c/p\u003e \u003cp\u003eSubgroup comparison revealed slightly weaker calibration performance in the mandible compared with the maxilla (bias: +2.3% vs\u0026thinsp;+\u0026thinsp;6.8%; JSD: 0.006 vs 0.004). Sex-based differences were minimal overall, though male patients demonstrated modestly larger over-prediction tendencies (21.0% vs 13.1%) despite similar rank correlations (ρ\u0026thinsp;=\u0026thinsp;0.726 vs 0.800).\u003c/p\u003e \u003cp\u003e Collectively, the digital setup provided a clinically acceptable representation of postoperative OGE prevalence and severity distribution, with prediction deviations largely within mild ranges across major subgroups.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCalibration and distribution agreement of OGEs prediction by subgroup.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGroup\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003en\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eObserved\u003c/p\u003e \u003cp\u003eincidence\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePredicted\u003c/p\u003e \u003cp\u003eincidence\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eBias\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eJSD\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSpearman_rho\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eExact\u003c/p\u003e \u003cp\u003ematch\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eUnder\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eOver\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e620\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e71.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e76.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.786\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e80.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e4.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e14.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMaxilla\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e310\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e57.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e64.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e6.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.740\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e82.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e4.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e12.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMandible\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e310\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e86.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e88.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.731\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e79.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e5.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e15.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e520\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e73.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e76.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e3.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.800\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e81.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e5.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e13.1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e64.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e75.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e11.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.726\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e75.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e4.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e21.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Binary prediction performance for OGE occurrence\u003c/h2\u003e \u003cp\u003eAs shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, binary prediction demonstrated high overall performance, with an accuracy of 91.6%, balanced sensitivity and specificity (0.916 and 0.918), and a high PPV (97.3%) with lower NPV (77.0%). The F1-score was 0.943.\u003c/p\u003e \u003cp\u003eMandibular sites showed higher accuracy and sensitivity than maxillary sites, which exhibited slightly more missed detections, while sex-based differences were modest, with higher sensitivity in females and higher specificity in males.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBinary performance metrics for predicting the presence of OGEs.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGroup\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003en\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eYouden_J\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e620\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.916\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.916\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.918\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.973\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.770\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.943\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMaxilla\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e310\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.874\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.850\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.918\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.950\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.771\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.897\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.768\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMandible\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e310\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.958\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.964\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.917\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.989\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.767\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.976\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.880\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e520\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.925\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.930\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.909\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.971\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.797\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.950\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.839\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.870\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.840\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.960\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.984\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.906\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.800\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eConfusion matrices (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), calibration plots (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), and ROC curves (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e; AUCs\u0026thinsp;\u0026gt;\u0026thinsp;0.8) confirmed stable and well-calibrated binary prediction across subgroups.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Agreement in OGE severity classification (ordinal Jemt grades)\u003c/h2\u003e \u003cp\u003eAs shown in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, predicted and actual Jemt grades demonstrated moderate-to-substantial agreement (κ\u0026thinsp;=\u0026thinsp;0.716\u0026ndash;0.774) with a low MAE (0.194). Exact-grade agreement was achieved in 80.8% of sites, and 99.8% of discrepancies were within \u0026plusmn;\u0026thinsp;1 grade.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eOverall and subgroup agreement between predicted and actual OGE severity (Jemt grades 0\u0026ndash;3).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGroup\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003en\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTop-1\u003c/p\u003e \u003cp\u003eaccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTop-1\u0026thinsp;\u0026plusmn;\u0026thinsp;1\u003c/p\u003e \u003cp\u003eaccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMAE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eWeighted\u0026nbsp;Kappa\u003c/p\u003e \u003cp\u003e(linear)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eWeighted\u0026nbsp;Kappa\u003c/p\u003e \u003cp\u003e(quadratic)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eSpearman\u0026nbsp;rho\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e620\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.808\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.998\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.194\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.716\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.786\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMaxilla\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e310\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.826\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.997\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.177\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.689\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.716\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.740\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMandible\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e310\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.790\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.210\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.680\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.743\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.731\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e520\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.819\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.181\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.733\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.790\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.800\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.750\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.260\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.633\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.695\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.726\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAgreement was slightly lower in mandibular sites and in male patients, although Spearman\u0026rsquo;s correlations remained strong across subgroups (ρ\u0026thinsp;\u0026ge;\u0026thinsp;0.726). Ordered confusion matrices (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e) showed predictions clustered along the diagonal, indicating predominantly minor misgrading.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Anatomical site\u0026ndash;dependent prediction performance\u003c/h2\u003e \u003cp\u003eAs shown in Tables\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and \u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, prediction performance varied modestly by papilla site. Binary accuracy exceeded 85% at most anterior sites, with greater stability in maxillary central and canine-adjacent embrasures and more variability in mandibular sites. Severity misclassification was predominantly minor (\u0026plusmn;\u0026thinsp;1 grade).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSite-specific binary prediction performance for postoperative OGEs presence.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSite\u0026nbsp;type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003en\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eYouden\u0026rsquo;s\u0026nbsp;J\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e13\u0026thinsp;\u0026minus;\u0026thinsp;12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.903\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.960\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.865\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.828\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.970\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.889\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.825\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12\u0026thinsp;\u0026minus;\u0026thinsp;11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.839\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.976\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.550\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.820\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.917\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.891\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.526\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e11\u0026ndash;21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.903\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.915\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.867\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.956\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.765\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.935\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.782\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e21\u0026ndash;22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.871\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.579\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.843\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.915\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.579\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e22\u0026ndash;23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.855\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.864\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.850\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.760\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.919\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.809\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.714\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e43\u0026thinsp;\u0026minus;\u0026thinsp;42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.935\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.979\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.786\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.940\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.917\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.959\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.765\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e42\u0026thinsp;\u0026minus;\u0026thinsp;41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.968\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.982\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.800\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.982\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.800\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.982\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.782\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e41\u0026thinsp;\u0026minus;\u0026thinsp;31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e31\u0026ndash;32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.919\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.982\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.429\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.931\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.750\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.956\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.410\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e32\u0026ndash;33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.968\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.962\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.980\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSite-specific agreement between predicted and actual OGEs severity (Jemt grades 0\u0026ndash;3).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSite\u0026nbsp;type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003en\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTop-1\u003c/p\u003e \u003cp\u003eaccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTop-1\u0026nbsp;\u0026plusmn;1\u003c/p\u003e \u003cp\u003eaccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMAE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eWeighted\u0026nbsp;Kappa\u003c/p\u003e \u003cp\u003e(linear)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eWeighted\u0026nbsp;Kappa\u003c/p\u003e \u003cp\u003e(quadratic)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eSpearman\u0026nbsp;rho\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e13\u0026thinsp;\u0026minus;\u0026thinsp;12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.903\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.097\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.816\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.828\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.819\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12\u0026thinsp;\u0026minus;\u0026thinsp;11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.694\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.306\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.472\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.559\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.604\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e11\u0026ndash;21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.871\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.129\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.744\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.773\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e21\u0026ndash;22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.839\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.161\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.631\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.656\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.696\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e22\u0026ndash;23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.823\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.984\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.194\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.628\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.613\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.679\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e43\u0026thinsp;\u0026minus;\u0026thinsp;42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.839\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.161\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.684\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.729\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.761\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e42\u0026thinsp;\u0026minus;\u0026thinsp;41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.694\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.306\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.527\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.610\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.576\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e41\u0026thinsp;\u0026minus;\u0026thinsp;31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.726\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.274\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.580\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.580\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e31\u0026ndash;32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.758\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.242\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.625\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.694\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.697\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e32\u0026ndash;33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.935\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.065\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.854\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.870\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.877\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eSite-specific confusion matrices showed stronger agreement in maxillary anterior regions (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), whereas mandibular central\u0026ndash;lateral sites exhibited greater dispersion (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). Overall, digital setup\u0026ndash;based site-specific prediction was clinically reliable, while highlighting mandibular incisor regions as requiring closer esthetic monitoring.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. DISCUSSION","content":"\u003cp\u003eThis retrospective study evaluated whether digital orthodontic setups, originally designed for clear aligner planning, can serve as a predictive tool for postoperative open gingival embrasure (OGE) development in adult extraction cases treated with fixed appliances. The results demonstrated that digital setups provided clinically acceptable prediction accuracy for both the presence and severity of papilla deficiency. Binary performance remained high across subgroups, with strong sensitivity indicating reliable detection of sites at risk. Severity estimates based on the modified Jemt index achieved substantial ordinal agreement, with nearly all discrepancies limited to minor\u0026thinsp;\u0026plusmn;\u0026thinsp;1\u0026ndash;grade differences. Site-based patterns highlighted the mandibular anterior region as the area with the greatest variability, consistent with its biomechanical vulnerability. Collectively, these findings support the potential of digital setups to assist clinicians in anticipating soft-tissue esthetic risks during treatment planning.\u003c/p\u003e \u003cp\u003eThe predictive value of digital orthodontic setups for OGE development aligns with established morphologic determinants of papilla loss. Prior research demonstrates that an ICP\u0026ndash;AC distance exceeding 5 mm, small increases in vertical discrepancies (ΔICP\u0026ndash;AC, ΔCEJ\u0026ndash;AC), and movement thresholds such as excessive intrusion or labial inclination significantly elevate OGE risk[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Because these geometric relationships\u0026mdash;contact-point position, crown form, root alignment, and interproximal spacing\u0026mdash;are inherently embedded within digital setups, the simulated final dentition can approximate the anatomical configurations most susceptible to papillary insufficiency, even without explicit soft-tissue modeling.\u003c/p\u003e \u003cp\u003eFrom a biological point, adults undergoing orthodontic treatment are inherently more susceptible to papillary recession because of age-related alveolar remodeling and diminished regenerative capacity, as papilla height decreased 0.012 mm each year of increasing age according to a published research[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Extraction therapy and crowding further exacerbate this vulnerability by requiring substantial tooth movement, root realignment, and space closure. As shown in previous evidences, the prevalence of OGEs is relatively high and strongly associated with tooth extraction, with severity decreasing in the order of one-lower-incisor extraction\u0026thinsp;\u0026gt;\u0026thinsp;two-premolar extraction\u0026thinsp;\u0026gt;\u0026thinsp;non-extraction cases[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Meanwhile, greater mandibular crowding was identified as an independent risk factor for OGE occurrence (OR\u0026thinsp;=\u0026thinsp;0.846) in particular[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. The anterior region, especially the mandibular incisors, is anatomically predisposed to OGE due to thin labial bone plates, limited keratinized tissue, and high sensitivity to even minor changes in root positioning[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. These biological tendencies are fully reflected in our study population\u0026mdash;adult extraction cases with \u0026ge;\u0026thinsp;4 mm crowding\u0026mdash;and correspond with the high overall OGE prevalence observed. Moreover, the site-dependent differences detected in our measurements, with mandibular anterior sites showing greater susceptibility, reinforce the established notion that local anatomy strongly influences papillary stability.\u003c/p\u003e \u003cp\u003eSeveral clinical studies have suggested that clear aligner therapy is associated with a higher risk of open gingival embrasures than fixed appliances. Yang et al. reported that the incidence of OGEs between maxillary and mandibular central incisors was 35.0% and 38.0% in the clear-aligner group, compared with 18.0% and 24.0% in the fixed-appliance group[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Cui et al. further showed that in non-extraction Invisalign cases the overall anterior OGE incidence reached 13.4% in the maxilla and 30.7% in the mandible, with the highest rate (\u0026asymp;\u0026thinsp;39%) between mandibular central incisors, and identified age, mandibular crowding and the distance from the interproximal contact point to the alveolar crest as independent risk factors[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. These findings are consistent with earlier work on fixed appliances[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], which emphasized the roles of anterior crowding, root divergence and contact-point\u0026ndash;bone relationships, but they also highlight that the \u0026ldquo;magnitude\u0026rdquo; and \u0026ldquo;frequency\u0026rdquo; of OGEs appear greater in clear-aligner cohorts.\u003c/p\u003e \u003cp\u003eOur results mirror this pattern at the level of digital prediction. Across the whole sample, the digitally generated setups systematically predicted a higher incidence and greater severity of OGEs than were actually observed after fixed-appliance treatment, both in overall analyses and within arch- and sex-stratified subgroups. Site-specific comparisons showed that predicted OGEs were particularly over-represented in mandibular anterior sites\u0026mdash;especially between the central incisors and adjacent contacts\u0026mdash;where clinical OGEs did occur but at a lower frequency. In terms of error structure, false-positive predictions clearly outnumbered false negatives in most subgroups, indicating that the digital setup behaves as a conservative \u0026ldquo;over-screening\u0026rdquo; tool for black triangles rather than an exact replica of the final soft-tissue outcome. This tendency is fully coherent with the aligner literature, in which mandibular anterior embrasures are consistently identified as the most vulnerable sites.\u003c/p\u003e \u003cp\u003eThe overestimation pattern may be partially explained by how aligner-based digital setups and algorithms encode risk. As discussed by previous studies[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], clear aligner systems often involve extensive labial and vertical movements, multiple attachments and prolonged coverage of the gingival margin; the aligner material extends apical to the contact point and may occupy the interdental space during papilla remodeling. These appliance-specific features, together with high ICP\u0026ndash;ABC distances and triangular crown forms, are implicitly built into the virtual setup and may bias the learned prediction models toward expecting more papillary deficiency than actually occurs under fixed-appliance biomechanics. In our study, however, all patients were treated with conventional fixed appliances, which may allow slightly better papilla preservation in some sites, leading to a gap between \u0026ldquo;virtual\u0026rdquo; and real soft-tissue outcomes. This bias towards overprediction has two important clinical implications. First, it suggests that setup-based OGE estimates should be understood as \u0026ldquo;upper-bound risk scenarios\u0026rdquo;: a high predicted probability or grade of OGE flags sites that truly deserve attention, but a proportion of these will not develop clinically visible black triangles. Second, the discrepancy between predicted and observed outcomes underscores the need to further refine current algorithms\u0026mdash;ideally by training on large, appliance-specific datasets and by explicitly incorporating known risk modifiers such as extraction pattern, crowding resolution strategy and periodontal phenotype. Until then, clinicians should interpret digital OGE predictions qualitatively, integrating them with clinical judgement.\u003c/p\u003e \u003cp\u003eSeveral limitations should be acknowledged when interpreting the present findings. First, this study adopted a retrospective design with data collected from a single clinical center, which may limit generalizability across populations with different periodontal phenotypes or treatment protocols. The sample primarily included adult extraction cases with moderate crowding; therefore, outcomes may differ in non-extraction or adolescent cohorts where papillary resilience and bone morphology vary. Second, although the predictive setup was generated by trained technicians using standardized digital protocols, the virtual alignment still lacks direct modeling of soft-tissue elasticity, gingival thickness, and alveolar crest remodeling\u0026mdash;factors that could influence papilla fill in vivo. Third, the Jemt index, while widely used and reproducible, remains a semi-quantitative visual scale and may not fully capture three-dimensional papilla morphology or subtle contour discrepancies[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Integration with volumetric soft-tissue analysis or digital photogrammetry may enhance assessment precision in future research.\u003c/p\u003e \u003cp\u003eTraditional methods for predicting the risk of open gingival embrasures rely on extensive clinical data collection\u0026mdash;patient age, crown morphology, alveolar and gingival biotype, root length and parallelism, tooth-movement patterns, periodontal probing measurements, and geometric calculations of contact-point\u0026ndash;crest relationships. Although informative, these approaches are analytically complex, time-consuming, and often unintuitive for both clinicians and patients. In contrast, digital orthodontic setups offer a data-driven alternative: machine-learning and deep-learning algorithms can synthesize large amounts of geometric and morphological information and output visually interpretable predictions directly from the \u0026ldquo;black box.\u0026rdquo; A scoping review reported that AI algorithms already achieve near-expert performance in tooth segmentation, CBCT\u0026ndash;IOS registration, and digital setup prediction, and can provide automated, visually interpretable outputs that support treatment planning and remote monitoring[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Similar conclusions were drawn in several reviews[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], which highlight the potential of AI to enhance efficiency and predictive accuracy in orthodontics. Against this backdrop, our study can be viewed as a posterior validation of the soft-tissue predictions generated by a commercial clear-aligner platform, demonstrating that digitally derived papilla simulations and OGE severity estimates align reasonably well with real clinical outcomes. These findings underscore the potential of AI-enhanced digital setups in soft-tissue outcome prediction. Future research should incorporate a more comprehensive set of established OGE risk factors to build advanced machine-learning models, and aligner companies may further refine their digital setup interfaces to improve the accuracy of soft-tissue rendering and prognostic reliability.\u003c/p\u003e \u003cp\u003eOverall, while the current approach demonstrates promising feasibility for soft-tissue risk visualization in fixed orthodontics, its translation to routine clinical practice requires further validation, model refinement, and standardization across different digital platforms.\u003c/p\u003e"},{"header":"5. CONCLUSION","content":"\u003cp\u003eDigital orthodontic setups derived from pre-treatment intraoral scans offer a feasible adjunct for predicting postoperative OGEs in adult extraction patients treated with fixed appliances, supporting their role in soft-tissue esthetic risk assessment and treatment planning.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAll authors contributed to the study conception and design. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003eAUTHOR CONTRIBUTIONS\u003c/p\u003e\n\u003cp\u003eXinyi, Ren: Methodology, Investigation, Visualization, Writing \u0026ndash; original draft.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eJiaxin, Deng: Data curation, Formal analysis, Software.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eLang Lei: Conceptualization, Writing \u0026ndash; review \u0026amp; editing.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eJialing, Li: Writing \u0026ndash; review \u0026amp; editing, Validation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHuang Li: Supervision, Project administration.\u003c/p\u003e\n\u003cp\u003eACKNOWLEDGMENTS\u003c/p\u003e\n\u003cp\u003eThis research was supported by Nanjing Health Development Key Project (ZKX23055). The authors thank the orthodontic technicians for their assistance with digital setup preparation. No external professional writing, editing, or proofreading services were used for this manuscript.\u003c/p\u003e\n\u003cp\u003eEthics approval \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis retrospective study was approved by the Ethics Committee of Nanjing Stomatological Hospital, Nanjing University (Approval No. NJSH-2023NL-036).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eConsent to participate \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe requirement for informed consent was waived by the Ethics Committee of Nanjing Stomatological Hospital, Nanjing University (Approval No. NJSH-2023NL-036) due to the retrospective design of the study.\u003c/p\u003e\n\u003cp\u003eConsent to publish\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003eConflict of Interest\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no conflict of interest.\u003c/p\u003e\n\u003cp\u003eFunding\u003c/p\u003e\n\u003cp\u003eThis work was supported by Nanjing Health Development Key Project (ZKX23055).\u003c/p\u003e\n\u003cp\u003eData availability\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003eAuthor contributions\u003c/p\u003e\n\u003cp\u003eXinyi, Ren: Methodology, Investigation, Visualization, Writing \u0026ndash; original draft. Jiaxin, Deng: Data curation, Formal analysis, Software. Lang Lei: Conceptualization, Writing \u0026ndash; review \u0026amp; editing. Jialing, Li: Writing \u0026ndash; review \u0026amp; editing, Validation. Huang Li: Supervision, Project administration.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eJ R, K. \u0026amp; V G, K. Open gingival embrasures after orthodontic treatment in adults: prevalence and etiology. Am J Orthod Dentofacial Orthop 120 (2001). https://doi.org/10.1067/mod.2001.114831\u003c/li\u003e\n\u003cli\u003eZhang, Y. et al. Clear aligners and open gingival embrasures: Retrospective study of epidemiology and risk factors. J Periodontol (2025). https://doi.org/10.1002/jper.11373\u003c/li\u003e\n\u003cli\u003eAn, S. S., Choi, Y. J., Kim, J. Y., Chung, C. J. \u0026amp; Kim, K. H. Risk factors associated with open gingival embrasures after orthodontic treatment. Angle Orthod 88, 267\u0026ndash;274 (2018). https://doi.org/10.2319/061917-399.12\u003c/li\u003e\n\u003cli\u003eZhang, Y. et al. IPR treatment and attachments design in clear aligner therapy and risk of open gingival embrasures in adults. Prog Orthod 24, 1 (2023). https://doi.org/10.1186/s40510-022-00452-1\u003c/li\u003e\n\u003cli\u003eErkang, T. et al. Factors influencing open gingival embrasures in orthodontic treatment: a retrospective clinical study. Prog Orthod 26 (2025). https://doi.org/10.1186/s40510-025-00554-6\u003c/li\u003e\n\u003cli\u003eSanabel O, B. Interdental papilla recession and reconstruction of the lost triangle: a review of the current literature. Front Dent Med 5 (2025). https://doi.org/10.3389/fdmed.2024.1537452\u003c/li\u003e\n\u003cli\u003eMahmoud K, A.-O. et al. The knowledge regarding the impacts and management of black triangles among dental professionals and laypeople. Sci Rep 14 (2024). https://doi.org/10.1038/s41598-024-61356-0\u003c/li\u003e\n\u003cli\u003eYudai, O. et al. Interdisciplinary Management of Open Interproximal Embrasures. J Esthet Restor Dent (2025). https://doi.org/10.1111/jerd.70015\u003c/li\u003e\n\u003cli\u003eYubohan, Z. et al. Biomechanical factors in the open gingival embrasure region during the intrusion of mandibular incisors: A new model through finite element analysis. Front Bioeng Biotechnol 11 (2023). https://doi.org/10.3389/fbioe.2023.1149472\u003c/li\u003e\n\u003cli\u003eChow, Y. C., Eber, R. M., Tsao, Y. P., Shotwell, J. L. \u0026amp; Wang, H. L. Factors associated with the appearance of gingival papillae. J Clin Periodontol 37, 719\u0026ndash;727 (2010). https://doi.org/10.1111/j.1600-051X.2010.01594.x\u003c/li\u003e\n\u003cli\u003eJung, J. S., Lim, H. K., Lee, Y. S. \u0026amp; Jung, S. K. The Occurrence and Risk Factors of Black Triangles Between Central Incisors After Orthodontic Treatment. Diagnostics (Basel) 14 (2024). https://doi.org/10.3390/diagnostics14232747\u003c/li\u003e\n\u003cli\u003eCui, W., Liu, Y., Zhao, Y., Lei, L. \u0026amp; Li, H. Risk factors for open gingival embrasures after clear aligners treatment: a retrospective study. BMC Oral Health 25, 547 (2025). https://doi.org/10.1186/s12903-025-05915-5\u003c/li\u003e\n\u003cli\u003eKim, T., Miyamoto, T., Nunn, M. E., Garcia, R. I. \u0026amp; Dietrich, T. Root proximity as a risk factor for progression of alveolar bone loss: the Veterans Affairs Dental Longitudinal Study. J Periodontol 79, 654\u0026ndash;659 (2008). https://doi.org/10.1902/jop.2008.070477\u003c/li\u003e\n\u003cli\u003eMahasneh, S. A., Goodwin, M., Pretty, I. \u0026amp; Cunliffe, J. The use of radiographs to assess the impact of the distance between the contact area and the crest of the bone to predict the presence or absence of interdental papilla: an in vivo study. Br Dent J (2023). https://doi.org/10.1038/s41415-023-6184-z\u003c/li\u003e\n\u003cli\u003eGuo, F. et al. Invisalign ClinCheck can predict open gingival embrasures in adult extraction cases: a pilot study. Angle Orthod 95, 389\u0026ndash;396 (2025). https://doi.org/10.2319/091424-752.1\u003c/li\u003e\n\u003cli\u003eLindsay, R. et al. Effectiveness of clear aligner therapy for orthodontic treatment: A systematic review. Orthod Craniofac Res 23 (2019). https://doi.org/10.1111/ocr.12353\u003c/li\u003e\n\u003cli\u003eLinghuan, R. et al. The predictability of orthodontic tooth movements through clear aligner among first-premolar extraction patients: a multivariate analysis. Prog Orthod 23 (2022). https://doi.org/10.1186/s40510-022-00447-y\u003c/li\u003e\n\u003cli\u003eGabriele, R., Simone, P., Tommaso, C., Andrea, D. \u0026amp; Cesare L, D. Efficacy of clear aligners in controlling orthodontic tooth movement: a systematic review. Angle Orthod 85 (2014). https://doi.org/10.2319/061614-436.1\u003c/li\u003e\n\u003cli\u003eTommaso, C. et al. Predictability of orthodontic tooth movement with aligners: effect of treatment design. Prog Orthod 24 (2023). https://doi.org/10.1186/s40510-022-00453-0\u003c/li\u003e\n\u003cli\u003eAbdalrahman Mohieddin, K., Kinda, S., Mohammad Younis, H. \u0026amp; Nikolaos, G. Digital setup accuracy for moderate crowding correction with fixed orthodontic appliances: a prospective study. Prog Orthod 25 (2024). https://doi.org/10.1186/s40510-024-00513-7\u003c/li\u003e\n\u003cli\u003eTianrui, Y. et al. The incidence and severity of open gingival embrasures in adults treated with clear aligners and fixed appliances: a retrospective cohort study. Head Face Med 19 (2023). https://doi.org/10.1186/s13005-023-00375-0\u003c/li\u003e\n\u003cli\u003eTadataka, I., Masaru, Y., Daijiro, M. \u0026amp; Kazutaka, K. Prediction and causes of open gingival embrasure spaces between the mandibular central incisors following orthodontic treatment. Aust Orthod J 20 (2006). \u003c/li\u003e\n\u003cli\u003eDaniele, C., Stefania, R. \u0026amp; Giuseppe, C. The Papilla Presence Index (PPI): a new system to assess interproximal papillary levels. Int J Periodontics Restorative Dent 24 (2004). https://doi.org/10.11607/prd.00.0596\u003c/li\u003e\n\u003cli\u003eD\u0026eacute;bora Costa, R. et al. Unveiling the role of artificial intelligence applied to clear aligner therapy: A scoping review. J Dent 154 (2025). https://doi.org/10.1016/j.jdent.2025.105564\u003c/li\u003e\n\u003cli\u003eSanjeev B, K. et al. Developments, application, and performance of artificial intelligence in dentistry - A systematic review. J Dent Sci 16 (2021). https://doi.org/10.1016/j.jds.2020.06.019\u003c/li\u003e\n\u003cli\u003eAndrej, T., Veronika, K. \u0026amp; Ivan, V. Artificial Intelligence in Orthodontic Smart Application for Treatment Coaching and Its Impact on Clinical Performance of Patients Monitored with AI-TeleHealth System. Healthcare (Basel) 9 (2021). https://doi.org/10.3390/healthcare9121695\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"Open gingival embrasures, Orthodontic tooth movement, Digital orthodontic setup, Extraction orthodontics, Soft tissue esthetics.","lastPublishedDoi":"10.21203/rs.3.rs-8439148/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8439148/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eObjectives\u003c/p\u003e\n\u003cp\u003eTo evaluate whether digital orthodontic setups can predict the presence and severity of postoperative open gingival embrasures (OGEs) in adult extraction cases treated with fixed appliances.\u003c/p\u003e\n\u003cp\u003eMaterials and Methods\u003c/p\u003e\n\u003cp\u003eThis retrospective study included 62 adults treated with four-premolar extractions and fixed appliances (620 anterior interproximal sites). Actual OGEs were assessed on post-treatment intraoral photographs. Predicted OGEs were assessed on digital setups generated on a clear aligner platform. Binary performance was evaluated using accuracy, sensitivity, specificity, PPV, NPV, F1 score, and Cohen’s κ. Severity agreement (modified Jemt grades) was assessed using weighted κ, mean absolute error (MAE), and Spearman’s ρ. Analyses were performed overall and by arch, sex, and site.\u003c/p\u003e\n\u003cp\u003eResults\u003c/p\u003e\n\u003cp\u003ePredicted OGE incidence approximated the observed rate (76.5% vs 71.9%; bias +4.5%). Binary prediction showed high performance (accuracy 91.6%, sensitivity 0.916, specificity 0.918, PPV 97.3%, NPV 77.0%, F1 0.943, κ 0.823). Severity prediction showed substantial agreement (weighted κ 0.716–0.774; MAE 0.194; ρ 0.786), with exact matches in 80.8% of sites and ±1-grade agreement in 99.8%.\u003c/p\u003e\n\u003cp\u003eConclusions\u003c/p\u003e\n\u003cp\u003eDigital setups provide clinically meaningful prediction of postoperative OGEs after fixed-appliance extraction treatment, with modest overprediction.\u003c/p\u003e\n\u003cp\u003eClinical Relevance\u003c/p\u003e\n\u003cp\u003eDigital setup–based prediction enables early identification of embrasure sites at higher esthetic risk, supporting proactive soft-tissue risk assessment and informed treatment planning in adult extraction orthodontics.\u003c/p\u003e","manuscriptTitle":"Digital Orthodontic Setup as a Predictive Tool for Post-treatment Open Gingival Embrasures in Adult Extraction Cases with Fixed Appliances: A Retrospective Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-12 16:03:19","doi":"10.21203/rs.3.rs-8439148/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":"f66554bd-4930-4ef0-b678-b1ab75e0a910","owner":[],"postedDate":"January 12th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-09T01:09:15+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-12 16:03:19","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8439148","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8439148","identity":"rs-8439148","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}