Arda Arısan

Objectives: To develop and validate an automated pipeline for maxillary dental midline localization on posed smile photographs and compare the performance of four convolutional neural network (CNN) backbones. Materials and Methods: A total of 221 standardized frontal posed smile photographs were collected. The maxillary dental midline was annotated by two orthodontists across two sessions, and mean coordinates served as the orthodontist-derived reference standard. An automated pipeline including face detection, interpupillary-line rotational normalization, mouth region extraction, and CNN-based keypoint regression was developed. Four CNN models (EfficientNet-B0, ResNet-18, MobileNetV3-Small, and HRNet-W18) were trained and compared using 5-fold stratified cross-validation. Deviation was expressed as the angle between the predicted dental midline and the interpupillary-derived photographic reference line (T-line). Results: HRNet-W18 achieved the lowest mean angular error (0.63° ± 0.63°) and the highest ≤1° error-threshold accuracy (81.0%). All 4 models exceeded 90% accuracy at the ≤2° threshold. No significant effect of tooth visibility grade on angular error was observed. HRNet-W18 also showed the highest agreement with the reference standard [intraclass correlation coefficient (ICC) = 0.721] and minimal Bland–Altman bias (+0.03°). Post hoc pairwise comparisons showed that HRNet-W18 outperformed all other models across all error metrics. Limitations: The sample was limited to standardized posed smile photographs from a single center, and generalizability to diverse populations and less controlled conditions remains to be established. Conclusions and Implications: The automated pipeline showed high accuracy for dental midline localization on posed smile photographs relative to an orthodontist-derived reference standard. Further validation on larger and more diverse datasets is needed before clinical application.

Background: This study evaluates the accuracy and repeatability of 3D Gaussian Splatting (3DGS)‐generated 3D facial models by comparing anthropometric measurements derived from these models with those obtained using conventional 2D photogrammetry and direct anthropometry. Methods: In this cross‐sectional study, 58 adults (18–25 years) underwent direct anthropometry, standardised 2D photography with calibration, and 3DGS facial reconstruction using a smartphone‐based workflow. A predefined set of linear and angular facial measurements was recorded; each measurement was repeated three times by the same observer. Between‐method comparisons were performed using paired tests as appropriate. Pearson correlation coefficients were reported to describe association, whereas agreement and potential interchangeability were evaluated primarily using Bland–Altman bias and 95% limits of agreement (LoA); plots were additionally inspected for proportional bias and influential outliers. Results: 3DGS demonstrated small mean differences relative to direct anthropometry for several measurements, with generally narrower LoA than 2D photogrammetry. Correlations were typically higher for 3DGS than for 2D photogrammetry; however, Bland–Altman analyses indicated that agreement varied by variable. For selected measurements, LoA were wide (reaching approximately 10–18 mm for some linear outcomes and exceeding 20° for certain angular outcomes), indicating limited interchangeability for individual‐level clinical decisions. Perioral measures showed greater dispersion for 2D photogrammetry, particularly for N–Stm, and a small number of observations fell outside the 95% LoA in ANT vs. 2D comparisons, consistent with occasional landmark identification or reconstruction‐related variability. Conclusion: 3D Gaussian Splatting (3DGS) may offer a useful alternative workflow for selected facial analysis applications by enabling the generation of measurable three‐dimensional facial models, from which anthropometric variables can subsequently be derived. Although 2D photogrammetry is readily accessible and practical to implement, its inherent limitation in representing three‐dimensional anatomy may compromise measurement accuracy in certain clinical contexts. Overall, these findings indicate that 3DGS could support orthodontic practice and research as a 3D model‐generation technique; however, further validation across diverse patient groups, acquisition conditions, expanded landmark‐based geometric error metrics, and additional outcome measures is warranted before routine clinical adoption can be recommended.

Arda Arısan, Celal Genç, Gökhan Serhat Duran. Korean J Orthod 2026;56:207-20. https://doi.org/10.4041/kjod25.320

Background: Lingual orthodontic systems have recently advanced with the introduction of fully customized CAD/CAM-based designs featuring self-ligating (SL) mechanisms. This study aimed to evaluate the three-dimensional accuracy of a customized SL lingual system in reproducing digitally planned tooth positions. Methods: A total of 280 teeth were analyzed following treatment with a fully customized self-ligating lingual system (Harmony^®, Aso International Inc., Tokyo, Japan). Digital models obtained before treatment (T0), from the setup (TS), and after treatment (T1) were superimposed using a best fit algorithm in GOM Inspect. Tooth movements were quantified across seven biomechanically relevant parameters including tip, torque, rotation, buccolingual, mesiodistal, vertical, and overall displacement. Predicted and achieved movements were compared using paired t tests and Bland–Altman analysis. Results: The fully customized SL lingual appliance achieved an overall dentition accuracy of 92.1%. Mean accuracy for linear tooth movements was 94.5% ± 2.1% in the maxilla and 93.8% ± 2.5% in the mandible. For angular movements, mean accuracy was 90.8% ± 3.4% in the maxilla and 89.3% ± 3.9% in the mandible. The highest precision was observed in anterior teeth for mesiodistal (96.2%) and buccolingual (95.8%) movements, whereas the lowest accuracy occurred in rotational movements of the posterior segments (87.1%). No statistically significant differences were found between predicted and achieved movements for most parameters (p > 0.05). Conclusions: The fully customized SL lingual orthodontic system demonstrated high accuracy in reproducing digitally planned tooth movements, particularly in the anterior segments. Although accuracy was slightly lower in the posterior regions, the overall outcomes remained mechanically and clinically acceptable across all evaluated dimensions.

Background: Multimodal large language models (LLMs) are increasingly being explored as clinical support tools, yet their capacity for orthodontic biomechanical reasoning has not been systematically evaluated. This retrospective study assessed their ability to analyze treatment mechanics and explored their potential role in supporting orthodontic decision-making. Methods: Five publicly available models (GPT-o3, Claude 3.7 Sonnet, Gemini 2.5 Pro, GPT-4.0, and Grok) analyzed 56 standardized intraoral photographs illustrating a diverse range of active orthodontic force systems commonly encountered in clinical practice. Three experienced orthodontists independently scored the outputs across four domains—observation, interpretation, biomechanics, and confidence—using a 5-point scale. Inter-rater agreement and consistency were assessed, and statistical comparisons were made between models. Results: GPT-o3 achieved the highest composite score (3.34/5.00; 66.8%), significantly outperforming all other models. The performance ranking was followed by Claude (57.8%), Gemini (52.6%), GPT-4.0 (48.8%), and Grok (38.8%). Inter-rater reliability among the expert evaluators was excellent, with ICC values ranging from 0.786 (Confidence Evaluation) to 0.802 (Observation). Model self-reported confidence showed poor calibration against expert-rated output quality. Conclusions: Multimodal LLMs show emerging potential for assisting orthodontic biomechanical assessment. With expert-guided validation, these models may contribute meaningfully to clinical decision support across diverse biomechanical scenarios encountered in routine orthodontic care.

Background: Monocular Depth Estimation (MDE) is a computer vision approach that predicts spatial depth information from a single two-dimensional image. In orthodontics, where facial soft-tissue evaluation is integral to diagnosis and treatment planning, such methods offer new possibilities for obtaining sagittal profile information from standard frontal photographs. This study aimed to determine whether MDE can extract clinically meaningful information for facial profile assessment. Methods: Standardized frontal photographs and lateral cephalometric radiographs from 82 adult patients (48 Class I, 28 Class II, 6 Class III) were retrospectively analyzed. Three clinically relevant soft-tissue landmarks—Upper Lip Anterior (ULA), Lower Lip Anterior (LLA), and Soft Tissue Pogonion (Pog′)—were annotated on frontal photographs, while true vertical line (TVL) analysis from cephalograms served as the reference standard. For each case, anteroposterior (AP) relationships among the three landmarks were represented as ordinal rankings based on predicted depth values, and accuracy was defined as complete agreement between model-derived and reference rankings. Depth maps were generated using one vision transformer model (DPT-Large) and two CNN-based models (DepthAnything-v2 and ZoeDepth). Model performance was evaluated using accuracy, 95% confidence intervals, and effect size measures. Results: The transformer-based DPT-Large achieved clinically acceptable accuracy in 92.7% of cases (76/82; 95% CI: 84.8–97.3), significantly outperforming the CNN-based models DepthAnything-v2 (9.8%) and ZoeDepth (4.9%), both of which performed below the theoretical chance level (16.7%). Conclusions: Vision transformer-based Monocular Depth Estimation demonstrates the potential for clinically meaningful soft-tissue profiling from frontal photographs, suggesting that depth information derived from two-dimensional images may serve as a supportive tool for facial profile evaluation. These findings provide a foundation for future studies exploring the integration of depth-based analysis into digital orthodontic diagnostics.