Predicting gingival embrasure risk after invisible orthodontics using multimodal data and machine learning

Clinical Snapshot

Key Findings

• P Value: All seven identified predictors p < 0.05 on multivariate analysis; no significant baseline differences between training and validation sets (p > 0.05)

• Effect Size: Random forest model: training AUC = 0.849; validation AUC = 0.815. Key predictors: gingival thickness and interproximal alveolar bone height (highest SHAP values). Independent risk factors: percentage of bleeding on probing–positive sites, interproximal alveolar bone height, relative movement of adjacent teeth. Protective factors: gingival thickness, proximal contact area, interdental papilla height, buccal/lingual bone plate thickness

• Primary Outcome: Development of gingival embrasure (black triangle) following clear aligner therapy

• Nnt Or Sensitivity: Hazard ratio not applicable (binary logistic/ML outcome). AUC of 0.815 on internal validation indicates moderate-to-good discrimination. Sensitivity, specificity, and positive/negative predictive values not reported in abstract. Net benefit on decision curve analysis described as favourable but thresholds not specified.

• Confidence Interval: Training AUC 95% CI: 0.786–0.912; Validation AUC 95% CI: 0.720–0.910

Clinical Application

Seven predictor variables require integration of clinical periodontal examination (bleeding on probing), cone-beam CT or periapical radiography (alveolar bone height, bone plate thickness), digital scanning or ultrasound (gingival thickness), and digital treatment planning software (relative tooth movement, proximal contact area, interdental papilla height). This multimodal data collection is feasible in well-equipped orthodontic or periodontic practices but may be resource-intensive in general dental settings. No validated clinical tool or app is yet available for deployment. Clear aligner therapy (e.g., Invisalign, ClearCorrect) is widely used in Australian orthodontic and general dental practice and is not PBS-subsidised, representing a significant out-of-pocket cost for patients. Gingival embrasure (black triangle) is a recognised complication affecting aesthetics and patient satisfaction, and is a medicolegal concern. The RACGP and Australian Society of Orthodontists do not currently have specific guidelines for gingival embrasure risk prediction. CBCT use in Australian dental practice is regulated by the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA) and is not routine for all clear aligner cases, which may limit the feasibility of the bone measurement predictors in standard Australian practice. TGA oversight applies to any software deployed as a medical device for clinical decision support. External validation in Australian or Western populations would be required before clinical adoption. Adult patients with malocclusion undergoing clear aligner orthodontic therapy, particularly those with pre-existing periodontal risk factors (bleeding on probing, reduced alveolar bone height, thin gingival biotype)

Abstract

OBJECTIVE: To develop and validate a risk prediction model for gingival embrasures after clear aligner therapy using multimodal oral data. METHODS: A retrospective study of 340 patients (December 2022-June 2025) was randomly divided into training (n = 238) and validation (n = 102) sets (7:3). Univariate analysis, multivariate logistic regression, and least absolute shrinkage and selection operator regression were applied to identify independent risk factors. Three machine learning models - random forest (RF), logistic regression, and support vector machine - were constructed based on seven core variables. Model performance was assessed using area under the receiver operating characteristic curve (AUC), calibration curves, decision curve analysis, and Shapley Additive Explanations (SHAP) values for interpretability. RESULTS: No significant baseline differences existed between sets (p > 0.05). Seven indicators were identified (p < 0.05). Multivariate analysis confirmed percentage of bleeding on probing-positive sites, interproximal alveolar bone height, and relative movement of adjacent teeth at target site as independent risk factors, while gingival thickness, proximal contact area, interdental papilla height, and buccal/lingual bone plate thickness were protective factors (p < 0.05). The RF model performed best: training AUC = 0.849 (95% CI: 0.786-0.912), validation AUC = 0.815 (95% CI: 0.720-0.910), with good calibration and net benefit. SHAP analysis highlighted gingival thickness and interproximal alveolar bone height as key predictors. CONCLUSION: A risk prediction model for post-clear aligner gingival embrasures was successfully developed and validated using multimodal oral data, with RF as the optimal algorithm. The model exhibits good discrimination, calibration, and clinical utility, which can be used as an objective auxiliary tool for individualized risk prediction following clear aligner therapy and supplement traditional clinical empirical judgment.

References

1. 1.Wang, H., Shi, H., Fan, L., Sun, Z., Xiu, X., & Yuan, H. (2026). Predicting gingival embrasure risk after invisible orthodontics using multimodal data and machine learning. Acta Odontologica Scandinavica. https://doi.org/10.2340/aos.v85.46562