Introduction : This study systematically compared the performance of 6 large language models (LLMs) from different vendors (Claude 3.7 Sonnet, ChatGPT 4.5, Gemini Flash 2.0, Grok 3, DeepSeek V3, Qwen) and 3 groups of LLMs from the same vendor (ChatGPT 4.5 vs ChatGPT 4o Mini, DeepSeek V3 vs DeepSeek R1, Qwen vs Qwen QWQ) on the same orthodontic multiple-choice question bank. The aim was to evaluate the performance of LLMs across different types of orthodontic knowledge and reasoning tasks within an educational context. Methods : A question bank consisting of 285 orthodontic multiple-choice questions was constructed, categorized according to two classification systems (Chapter-based Classification and Examination Direction-based Classification). The models were then tested through a standardized process to compare their performance across vendors and within the same vendor. Results : Claude 3.7 Sonnet (88.4%) and ChatGPT 4.5 (87.0%) ranked first and second in overall accuracy. In category-wise performance, Claude 3.7 Sonnet led in Basic Knowledge (96.8%), Clinical Strategy (90.8%), Type A (92.0%), and Type B (90.8%). ChatGPT 4.5 excelled in Fundamental Theory (86.3%) and Type C (83.8%). Within the same vendor comparisons, ChatGPT 4.5 and Qwen significantly outperformed their counterparts. DeepSeek V3 and DeepSeek R1 showed comparable performance. Conclusions : Claude 3.7 Sonnet and ChatGPT 4.5 demonstrated the strongest overall performance in this evaluation. LLMs generally perform better on Clinical Strategy-related questions and tasks requiring moderate reasoning (Type B). However, performance differences among models from the same vendor are inconclusive, necessitating further research to validate LLM capabilities in orthodontics and provide empirical evidence for understanding LLM performance in orthodontic education.
Introduction Molar band selection during fixed orthodontic treatment has traditionally relied on trial-and-error, a process that is time-consuming, unhygienic, and difficult to standardize across operators. A predictable, reproducible method based on digital tooth measurements could improve efficiency and support the development of standardized clinical protocols. Methods A total of 128 Ormco orthodontic bands (Ormco Corporation, Glendora, Calif) were digitized to establish an averaged circumference reference table. Tooth circumference was measured from intraoral scans using an eight-point custom spline in OrthoAnalyzer™ software (3Shape A/S, Copenhagen, Denmark) at the level of the marginal ridges. Predicted band sizes were compared with clinically selected sizes in 106 first and second molars from 42 patients. Intra-operator reliability was assessed across four repeated measurement sessions. Results Exact agreement between predicted and clinical band size was achieved in 48.1% of cases, with agreement within ±1 size in 70.8% and within ±2 sizes in 95.3%. Descriptively, most prediction errors were positive. The lowest exact agreement was observed in mandibular first molars. The intraclass correlation coefficient (ICC) for repeated circumference measurements was 0.988. The standard error of measurement (0.281 mm) was below the mean increment between adjacent band sizes (0.403 ± 0.016 mm). Conclusions Digital tooth circumference measurement from intraoral scans provides a reproducible and clinically applicable method for predicting molar band size, supporting its integration into standardized orthodontic protocols.
This article reports the combined surgical exposure and orthodontic treatment of labial inversely impacted maxillary central incisors in the mixed and primary dentitions. After being successfully moved into proper position, one highly positioned impacted incisor had a short root with an obliterated root canal but showed an aesthetically good and stable outcome fourteen years posttreatment. The other two impacted incisors had a normal root morphology posttreatment, similar to the roots of their contralateral central incisors. The surgical procedures and the orthodontic strategies for treating labial inversely impacted maxillary central incisors are described with a review of the current literature on the topic. The importance of early treatment in the root development of labial inversely impacted maxillary central incisors is also discussed.
