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

Automated assessment of gingival biotype using deep learning on intraoral photographs

Gingival biotype is a key factor influencing dental treatment outcomes. This study aimed to construct and validate an artificial intelligence (AI) model for objective and reproducible gingival biotype assessment based on intraoral photographs, thereby supporting clinical decision-making and personalized treatment planning. A total of 1,600 participants (aged 24 ± 2 years; 720 males and 880 females) with healthy periodontal conditions were enrolled. Gingival biotype was clinically identified using the probe transparency method and categorized as thick, medium, or thin. The dataset (640 thick, 520 medium, 440 thin) was split into a training set ( n = 1,500) and testing set ( n = 100) with proportional distribution. Weighted cross-entropy was applied to account for class imbalance. The Vision Transformer (ViT) model was trained using AdamW with an initial learning rate of 0.001 and a batch size of 8 for 10 epochs, whereas Residual Network-18 (ResNet-18) was trained using Adam with a learning rate of 1 × 10 −4 and a batch size of 32 with early stopping. Offline data augmentation (rotation ±15°, horizontal/vertical flipping, contrast/gamma adjustment, and contrast-limited adaptive histogram equalization (CLAHE)) was applied in a 1:4 ratio with fixed random seeds (123) to expand the training set from 1,500 to 7,500 images, while the test set underwent fixed preprocessing only. Model performance was evaluated using F1-score and area under the receiver operating characteristic curve (AUC), with statistical significance and 95% confidence intervals estimated via bootstrap resampling and DeLong test. The ViT model achieved superior diagnostic performance compared with ResNet-18. In the test set, the ViT model reached AUCs of 1.00 (thick), 0.98 (thin), and 0.97 (medium), significantly higher than those of ResNet-18 (0.82, 0.79, and 0.80; p < 0.001). The ViT model also showed improved recall and F1-scores, particularly for the “Thick” and “Thin” classes, and better class separability in the confusion matrix, confirming its robustness and statistical reliability. The proposed deep learning model, particularly the Vision Transformer, demonstrated high diagnostic accuracy and robust performance across gingival biotype classes. The ViT model offers predictive capability and potential to support clinical decision-making in treatment planning. This artificial intelligence (AI)-based approach enables non-invasive, objective and efficient gingival biotype assessment, facilitating early risk evaluation and personalized treatment planning. Integrated into digital diagnostic workflows, it can assist clinicians in selecting appropriate incision designs, restorative margin levels, and orthodontic force strategies according to biotype characteristics, thereby improving treatment predictability and patient outcomes.

Predictive Analysis for Success and Complications in Dental Implant Therapy using Artificial Intelligence Models

Introduction: Dental implant therapy is a predictable treatment for tooth replacement, but its success is influenced by multiple patient, surgical, and systemic factors. Traditional statistical models, such as logistic regression, are limited in capturing nonlinear relationships among these variables. Artificial intelligence (AI) can integrate diverse predictors to enhance clinical risk assessment and improve treatment outcomes. Objective: To identify key predictors of dental implant success and develop AI-based models capable of accurately forecasting implant outcomes using clinical and surgical parameters. Methods: This retrospective cohort study analyzed data from 172 patients (219 implants) placed between 2020 and 2021. Patient demographics, systemic health (smoking, diabetes), surgical variables, and implant characteristics were evaluated. Univariate and multivariate logistic regression identified independent predictors. Machine-learning models Logistic Regression, Decision Tree, Support Vector Machine (SVM), and Random Forest were trained using stratified 10-fold cross-validation and SMOTE balancing. Performance was assessed via accuracy, precision, recall, F1-score, AUC, and Brier score; calibration and Decision-Curve Analysis (DCA) evaluated model reliability and clinical benefit. Results: Overall implant success was 91.3%, with smoking (AOR D 2.3; 95% CI 1.3–4.1; p D 0.001) and diabetes mellitus (AOR D 1.8; 95% CI 1.1–3.5; p D 0.03) emerging as independent predictors of failure. Flapless surgery demonstrated a protective effect (AOR D 0.7; 95% CI 0.5–0.9; p D 0.04). The Random Forest model achieved the highest predictive performance (Accuracy D 87.4%, AUC D 0.91) and showed good calibration (Brier D 0.07) with superior clinical net benefit on DCA. Conclusion: AI-based modeling offers a robust, data-driven approach for predicting dental implant success by integrating multifactorial clinical parameters. Incorporating such models into clinical workflows can enhance patient-specific risk assessment and improve long-term implant outcomes.

Machine learning, decision tree and nomogram for predicting screw loosening after PLIF in osteoporotic patients: a retrospective multicenter study

To develop and validate machine learning models and an individualized nomogram for predicting pedicle screw loosening after posterior lumbar interbody fusion (PLIF) in osteoporotic patients using preoperative clinical, imaging, and bone metabolism-related medication profiles. A retrospective analysis was conducted on 630 osteoporotic patients who underwent PLIF at three spine surgery centers. Patients were divided into a non-loosening group ( n = 450) and a loosening group ( n = 180) according to the presence of implant loosening on imaging within 12 months postoperatively. Univariate analysis was used to screen candidate variables, and LASSO–logistic regression with 10-fold cross–validation was applied to extract independent predictors. Four models (naïve Bayes, logistic regression, linear discriminant analysis, and decision tree) were constructed based on the selected features and evaluated using the area under the curve (AUC), calibration curves, and decision curve analysis (DCA). The structure of the optimal model (decision tree) was visualized, and a nomogram was built using multivariable logistic regression. Univariate analysis showed significant differences between the two groups in age, bone mineral density T-score, pelvic incidence, lumbar lordosis, sex, hypertension, diabetes mellitus, foraminal morphology, history of glucocorticoid use, calcium supplementation, and vitamin D supplementation (all P < 0.05). LASSO regression identified 11 independent predictors (λ.min = 0.0325). Among the four models, the decision tree showed the highest discrimination, achieving an AUC of 0.975 (95% CI 0.961–0.989) in the training set and 0.971 (95% CI 0.955–0.987) in the test set, with good calibration (Hosmer-Lemeshow test P = 0.418). DCA demonstrated a significant net benefit across a clinically relevant threshold probability range (0–50%). The decision tree identified the bone mineral density T-score as the root splitting variable; a T-score < –3.6 classified patients as being at extremely high risk for loosening. Among those with T-score ≥ –3.6 who did not take calcium, lack of vitamin D supplementation was associated with a very high risk. Among those with T-score ≥ –3.6 who took calcium, female sex combined with lumbar lordosis ≥ 51° indicated an elevated risk. The nomogram integrated the 11 factors and yielded a C-index of 0.928 (95% CI 0.903–0.953) with good calibration (Hosmer-Lemeshow test P = 0.372). The decision tree model demonstrated favorable predictive performance for pedicle screw loosening after PLIF. Its interpretable classification rules facilitate rapid screening of high-risk patients, while the nomogram enables individualized probability estimation for surgical planning. Together, these complementary tools offer a potentially useful framework for preoperative risk stratification and personalized management of osteoporotic patients undergoing PLIF

Assessing the Diagnostic Performance of ChatGPT-5.0 versus Machine Learning in Orthodontics: A Comparative Analysis for Extraction Treatment Planning.

Objective To make accurate orthodontic extraction decisions, various clinical and cephalometric variables must be evaluated. This study aims to evaluate ChatGPT-5.0's performance in distinguishing orthodontic extraction decisions and to compare it with five supervised machine learning (ML) algorithms. Methods Of 550 retrospectively evaluated orthodontic records, 30 were reserved for calibration, leaving 520 for the main analysis. The reference standard was the consensus treatment decision of three expert orthodontists with more than 5 years of clinical experience. Overall, 23 variables were analyzed, including 13 clinical parameters, 7 cephalometric measurements, and photographs. ChatGPT-5.0's performance was evaluated using a 5-fold cross-validation design. It was compared with XGBoost, random forest, support vector machine (SVM), logistic regression, and multi-layer perceptron (MLP). Performance metrics included accuracy, sensitivity, specificity, precision, F1-score, and balanced accuracy, with 95% confidence intervals calculated. Statistical analyses utilized Cochran's Q test and the McNemar test with Holm-Bonferroni correction. Results Of the 520 main cases, 223 (42.88%) were extraction treatments and 297 (57.12%) were non-extraction treatments. XGBoost achieved the highest accuracy (78.08%), followed closely by ChatGPT-5.0 (75.77%). The overall performance difference among models was significant (p≤0.001). In pairwise comparisons, ChatGPT's accuracy was significantly higher than those of random forest, SVM, logistic regression, and MLP, but was found to be similar to XGBoost. ChatGPT-5.0 showed the highest sensitivity (76.68%), whereas XGBoost showed the highest specificity (82.15%). Conclusion ChatGPT-5.0 demonstrated performance comparable to, and in some cases superior to, traditional ML models for orthodontic extraction decisions. While XGBoost yielded the highest overall classification accuracy, ChatGPT-5.0's high sensitivity in detecting extraction cases was noteworthy.

Predicting patient satisfaction with post-small incision lenticule extraction surgery using open-source machine learning software.

The effect of Sequential Retraction of Maxillary Incisors on Anterior Gingival Biotypes

Objective: To evaluate the effect of sequential retraction of maxillary incisors on the gingival biotypes Methodology: This cross-sectional was conducted in the Department of Orthodontics at Altamash Institute of Dental Medicine, Karachi. Duration of study was one year from June 2023 to July 2024. A total of 50 patients aged 18-30 years, diagnosed with Angle Class II Division 1 malocclusion and ANB > 4°, were included. Participants were categorized into two groups, thin (n=35) and thick (n=15) gingival biotypes. Gingival parameters were assessed before and after maxillary incisor retraction using a World Health Organization (WHO) probe also known as the Community Periodontal Index of Treatment Needs (CPITN) probe. Quantitative data, the gingival height, probing depth, and gingival width, were assessed using the Shapiro-Wilk test. Statistical analysis was performed using SPSS with level of significance set at p ≤ 0.05. Results: Average age of patients was 24.5 years, comprising 40% males and 60% females. Among them, 70% had a thin gingival biotype, and 30% had a thick biotype. Thin biotype patients had lower mean gingival thickness (0.88 mm) and width (3.2 mm) than thick biotype patients (1.25 mm and 3.5 mm, respectively). Probing depth and gingival recession were slightly higher in thin biotypes. THIn the thin biotype group, gingival thickness decreased by 0.06 mm, probing depth increased by 0.2 mm, gingival width reduced by 0.2 mm, and gingival recession rose by 0.2 mm. For the thick biotype, changes were less pronounced, with a decrease in gingival thickness by 0.03 mm, a slight increase in probing depth by 0.1 mm, a reduction in gingival width by 0.1 mm, and an increase in gingival recession by 0.1 mm. After sequential retraction of maxillary teeth patients with a thin biotype showed significant reductions in gingival thickness (p=0.03) and width (p=0.02) with an increase in probing depth (p=0.04) and gingival recession (p=0.01). In contrast, the thick biotype group demonstrated minimal and statistically insignificant changes in these parameters (p > 0.05). Conclusion: Patients with a thin biotype have increased risk of periodontal recession and reduction of gingival thickness while retraction of anterior teeth. Orthodontic treatments should consider appropriate measures to retract teeth keeping the biotype of gingiva in mind to preserve gingival health.