DOI 10.1016/j.ajodo.2025.09.009
Clinical advances in curve of Spee assessment: Deep learning for automatic tooth landmark detection in Invisalign.
Аннотация
INTRODUCTION
The curve of Spee (COS) is a key indicator of occlusal function and orthodontic outcomes. Its measurement traditionally relies on manual landmark identification from intraoral scans, which is time-consuming and operator-dependent. This study introduces a deep learning-based method for fully automated COS assessment based on intraoral scan data, aiming to improve measurement efficiency and support the evaluation of the predictability of COS leveling in patients with varying vertical skeletal patterns undergoing Invisalign treatment (Align Technology, Santa Clara, Calif).
METHODS
In this retrospective study, a total of 194 mandibular arch models were used to train and validate an automated network for measuring COS. This network adopted a Structure-Aware Long Short-Term Memory framework, which employed a 2-stage method for detecting coarse and fine tooth landmarks to assess COS depth. The accuracy of the landmarks was evaluated using the mean radial error and success detection rate, whereas COS depth was assessed using the paired Wilcoxon test. Finally, 55 patients with different vertical skeletal patterns were selected to analyze differences in COS leveling effects. In addition, the extrusion of teeth relative to the occlusal plane was compared.
RESULTS
The proposed network closely approximated the manual method, with the mean radial error for landmark detection being 0.32 ± 0.09 mm. The median measurement error for COS was 0.05 mm (P <0.001). ClinCheck predicted an average of 0.24 mm higher COS leveling than the actual outcome (P <0.01). The hyperdivergent group exhibited the highest predictability at 69%, whereas the hypodivergent group showed the lowest predictability at 58%. The accuracy of extrusion relative to the occlusal plane at the first molar was the lowest (65%).
CONCLUSIONS
Deep learning can aid in measuring COS. In ClinCheck, considering various vertical skeletal patterns is necessary when designing the leveling objectives for COS, with the first molar requiring particular attention.
Предметные рубрики MeSH
Подготовка и перевод публикации
Обработка статьи · этапы, ошибки и управление
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Тип публикации указан по данным источника. Уровень доказательности автоматически не присваивается.