Journal of the World Federation of Orthodontists最新文献_第8页

6 - A comparison of the accuracy of artificial intelligence versus orthodontic clinicians in identifying cephalometric landmarks 人工智能与正畸临床医生在识别头侧标志的准确性比较

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-12-09 DOI: 10.1016/j.ejwf.2025.07.010

Yen Ming Lin , Asa Auta , Robert Ireland , Richard Cure , Jeremy Brown

Background

Accurate identification of cephalometric landmarks is essential for orthodontic diagnosis and treatment planning. Manual landmarking by orthodontic clinicians can be time-consuming and variable, particularly due to differences in experience. To address this, artificial intelligence (AI) algorithms have been developed to automate cephalometric analysis, aiming to improve efficiency, accuracy, and consistency.

Objectives

This study aimed to evaluate the accuracy of selected skeletal and dental cephalometric landmarks and the resulting linear and angular measurements identified by commercial AI-based providers, compared to those obtained manually by orthodontic clinicians with varying experience against a reference standard.

Methods

Thirty-five lateral cephalometric radiographs from the IEEE ISBI Grand Challenge 2015 dataset were analysed. Ten landmarks were identified on each radiograph by nine calibrated orthodontic clinicians and four commercial AI providers. These were used to calculate nine cephalometric measurements, compared to a reference standard. Agreement among the clinicians was evaluated using the intraclass correlation coefficient (ICC). Friedman’s test, Wilcoxon signed ranks test with Bonferroni correction, and Bland-Altman plots with linear regression were used to test for differences between groups.

Results

ICC values showed six of nine measurements from human clinicians had excellent inter-rater reliability. However, Friedman’s test indicated significant differences between the human clinicians and AI providers against the reference standard for all measurements. Post hoc analysis confirmed these findings. Bland-Altman plots also showed reduced accuracy and consistency across groups.

Conclusions

Both AI algorithms and manual methods showed limited agreement with the reference standard. Further research is needed to improve the accuracy, reliability, and clinical use of AI in cephalometric analysis.

背景：准确识别头部测量标志对于正畸诊断和治疗计划至关重要。正畸临床医生手动标记可能是耗时和可变的，特别是由于经验的差异。为了解决这个问题，人们开发了人工智能（AI）算法来自动进行头测分析，旨在提高效率、准确性和一致性。本研究旨在评估由商业人工智能提供者识别的选定骨骼和牙齿头部测量标志以及由此产生的线性和角度测量的准确性，与具有不同经验的正畸临床医生根据参考标准手动获得的结果进行比较。方法分析来自IEEE ISBI Grand Challenge 2015数据集的35张侧位头颅x线片。9名校准的正畸临床医生和4名商业人工智能提供者在每张x光片上确定了10个地标。与参考标准相比，这些被用来计算9个头侧测量值。临床医生之间的一致性使用类内相关系数（ICC）进行评估。采用Friedman检验，采用Bonferroni校正的Wilcoxon符号秩检验，采用线性回归的Bland-Altman图检验组间差异。结果icc值显示，来自人类临床医生的9个测量值中有6个具有出色的评分间可靠性。然而，Friedman的测试表明，与所有测量的参考标准相比，人类临床医生和人工智能提供者之间存在显著差异。事后分析证实了这些发现。Bland-Altman图也显示了组间准确性和一致性的降低。结论人工智能算法和人工方法与参考标准的一致性有限。需要进一步研究以提高人工智能在头颅测量分析中的准确性、可靠性和临床应用。

{"title":"6 - A comparison of the accuracy of artificial intelligence versus orthodontic clinicians in identifying cephalometric landmarks","authors":"Yen Ming Lin , Asa Auta , Robert Ireland , Richard Cure , Jeremy Brown","doi":"10.1016/j.ejwf.2025.07.010","DOIUrl":"10.1016/j.ejwf.2025.07.010","url":null,"abstract":"<div><h3>Background</h3><div>Accurate identification of cephalometric landmarks is essential for orthodontic diagnosis and treatment planning. Manual landmarking by orthodontic clinicians can be time-consuming and variable, particularly due to differences in experience. To address this, artificial intelligence (AI) algorithms have been developed to automate cephalometric analysis, aiming to improve efficiency, accuracy, and consistency.</div></div><div><h3>Objectives</h3><div>This study aimed to evaluate the accuracy of selected skeletal and dental cephalometric landmarks and the resulting linear and angular measurements identified by commercial AI-based providers, compared to those obtained manually by orthodontic clinicians with varying experience against a reference standard.</div></div><div><h3>Methods</h3><div>Thirty-five lateral cephalometric radiographs from the IEEE ISBI Grand Challenge 2015 dataset were analysed. Ten landmarks were identified on each radiograph by nine calibrated orthodontic clinicians and four commercial AI providers. These were used to calculate nine cephalometric measurements, compared to a reference standard. Agreement among the clinicians was evaluated using the intraclass correlation coefficient (ICC). Friedman’s test, Wilcoxon signed ranks test with Bonferroni correction, and Bland-Altman plots with linear regression were used to test for differences between groups.</div></div><div><h3>Results</h3><div>ICC values showed six of nine measurements from human clinicians had excellent inter-rater reliability. However, Friedman’s test indicated significant differences between the human clinicians and AI providers against the reference standard for all measurements. Post hoc analysis confirmed these findings. Bland-Altman plots also showed reduced accuracy and consistency across groups.</div></div><div><h3>Conclusions</h3><div>Both AI algorithms and manual methods showed limited agreement with the reference standard. Further research is needed to improve the accuracy, reliability, and clinical use of AI in cephalometric analysis.</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Page 399"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705770","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

8 - A facial maturation index based on cranium-to-face ratio as a predictor of skeletal maturity 基于颅面比的面部成熟指数作为骨骼成熟度的预测指标

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-12-09 DOI: 10.1016/j.ejwf.2025.07.012

Gabriela Trojahn , Paula Balestrin , Kelly Chiqueto , Eduardo Ferreira , Katherine Jaramillo Cevallos , Sérgio Estelita Barros

Background

Diagnostic methods to determine individual skeletal maturity have been reported to indicate treatment timing in orthodontics.

Objective

To evaluate the changes in cranium/face size ratio (Facial Maturation Index - FMI) between the skeletal maturity stages defined by cervical vertebral maturation (CVM).

Methods

The sample consisted of a set of lateral cephalograms selected from the records of 180 subjects from the Burlington Growth Center and the Forsyth Institute. For each lateral cephalogram, the face area (FA) and neurocranium area (NCA) were delimited and measured. The FMI was calculated by FA/NCA. The FMI was compared and correlated with the CVM stages using ANOVA and Spearman's correlation tests. ROC analysis evaluated the FMI diagnostic performance.

Results

The NCA was similar between all maturation stages. FA and FMI increased significantly with each advance in CVM stages. The female sample tended to have a lower facial growth rate and earlier facial growth acceleration, while males tended to have a longer growth spurt period. There was significant and positive correlation between FMI and CVM stages for female (0.768) and male (0.803) samples. Areas under ROC curves for FMI were 0.851 to female and 0.854 to male sample, showing good diagnostic performance of FMI for discriminating pre-peak maturation stage from peak and post-peak stages.

Conclusion

Changes in cranium-to-face ratio were strongly correlated with CVM maturation stages. The diagnostic performance of FMI shows that it is a clinically useful diagnostic aid for predicting facial growth.

背景：据报道，确定个体骨骼成熟度的诊断方法可以指示正畸治疗的时机。目的评价以颈椎成熟（CVM）为指标的不同骨骼成熟阶段颅骨/面部大小比值（面部成熟指数- FMI）的变化。方法从伯灵顿生长中心和福赛斯研究所180名受试者的记录中选择一套侧位脑电图作为样本。对于每个侧位脑电图，面部面积（FA）和神经头盖骨面积（NCA）被划定和测量。FMI采用FA/NCA计算。使用方差分析和Spearman相关检验比较FMI与CVM分期的相关性。ROC分析评价FMI诊断效果。结果各成熟阶段的NCA相似。随着CVM分期的进展，FA和FMI显著增加。女性的面部生长速度较慢，面部生长加速较早，而男性的生长突增期较长。女性（0.768）和男性（0.803）的FMI与CVM分期呈显著正相关。女性FMI的ROC曲线下面积为0.851，男性FMI的ROC曲线下面积为0.854，表明FMI对峰前成熟阶段、峰后成熟阶段有较好的诊断作用。结论颅面比的变化与CVM的成熟阶段密切相关。FMI的诊断性能表明它是一种临床上有用的预测面部生长的诊断辅助工具。

{"title":"8 - A facial maturation index based on cranium-to-face ratio as a predictor of skeletal maturity","authors":"Gabriela Trojahn , Paula Balestrin , Kelly Chiqueto , Eduardo Ferreira , Katherine Jaramillo Cevallos , Sérgio Estelita Barros","doi":"10.1016/j.ejwf.2025.07.012","DOIUrl":"10.1016/j.ejwf.2025.07.012","url":null,"abstract":"<div><h3>Background</h3><div>Diagnostic methods to determine individual skeletal maturity have been reported to indicate treatment timing in orthodontics.</div></div><div><h3>Objective</h3><div>To evaluate the changes in cranium/face size ratio (Facial Maturation Index - FMI) between the skeletal maturity stages defined by cervical vertebral maturation (CVM).</div></div><div><h3>Methods</h3><div>The sample consisted of a set of lateral cephalograms selected from the records of 180 subjects from the Burlington Growth Center and the Forsyth Institute. For each lateral cephalogram, the face area (FA) and neurocranium area (NCA) were delimited and measured. The FMI was calculated by FA/NCA. The FMI was compared and correlated with the CVM stages using ANOVA and Spearman's correlation tests. ROC analysis evaluated the FMI diagnostic performance.</div></div><div><h3>Results</h3><div>The NCA was similar between all maturation stages. FA and FMI increased significantly with each advance in CVM stages. The female sample tended to have a lower facial growth rate and earlier facial growth acceleration, while males tended to have a longer growth spurt period. There was significant and positive correlation between FMI and CVM stages for female (0.768) and male (0.803) samples. Areas under ROC curves for FMI were 0.851 to female and 0.854 to male sample, showing good diagnostic performance of FMI for discriminating pre-peak maturation stage from peak and post-peak stages.</div></div><div><h3>Conclusion</h3><div>Changes in cranium-to-face ratio were strongly correlated with CVM maturation stages. The diagnostic performance of FMI shows that it is a clinically useful diagnostic aid for predicting facial growth.</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Page 399"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705771","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Long term stability of molar position postdistalization: A systematic review and meta-analysis 磨牙位置远端后的长期稳定性：系统回顾和荟萃分析。

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-09-23 DOI: 10.1016/j.ejwf.2025.07.734

Kutraaleeshwaran Velmurugan, Venkateswaran Ananthanarayanan, Sridevi Padmanabhan, Vignesh Kailasam

Background

Molar distalization has generated additional attention with the advent of skeletal anchorage devices. However, relapse after distalization is one of the challenges and the quantum of relapse following distalization, especially over the long-term remains controversial. This systematic review and meta-analysis aim to evaluate the long-term vertical and sagittal stability of molars postdistalization.

Methods

A systematic search was conducted in electronic databases (PubMed, Scopus, Ovid, Embase, Cochrane and LILACS) up to April 11, 2025. Studies assessing molar position changes after active distalization and during follow-up periods were included. Data extraction and risk of bias assessments were performed using the ROBINS-I tool for non-randomized studies. A meta-analysis was conducted to quantify relapse in vertical and sagittal dimensions.

Results

From the 2594 studies initially identified, six studies (all nonrandomized control trials) were included for the systematic review and five studies were included for the meta-analysis. All studies showed moderate risk of bias. The mean relapse in the sagittal plane was 1.58 ± 0.84 mm (35%) linearly and 2.98 ± 2.07° (53%) angularly. In the vertical plane, it relapsed by 1.44 ± 1.26 mm. Meta-analysis showed a net distalization in the molar position in long-term sagittal plane by 2.77 mm (95% CI 2.33–3.20) and 1° distally (95% CI −2.37° to 4.38°), and in the vertical plane by −0.65 mm (95% CI −2.27 to 0.96). The I² ranged from 60% to 98% indicating high heterogeneity. The relapse was minimal for the incisal, skeletal, and soft tissues parameters.

Conclusions

Long-term assessment following molar distalization demonstrated conclusive evidence of relapse in the sagittal and vertical direction. There was greater relapse in the vertical than in the sagittal direction. There was minimal literature assessing the relapse in the transverse direction. Hence, this systematic review advocates the need for long-term retention postdistalization.

背景：随着骨骼锚定装置的出现，磨牙远端引起了更多的关注。然而，远端术后复发是一个挑战，远端术后复发的数量，特别是长期的复发仍然存在争议。本系统综述和荟萃分析旨在评估磨牙远端后的长期垂直和矢状稳定性。方法：系统检索截至2025年4月11日的PubMed、Scopus、Ovid、Embase、Cochrane、LILACS等电子数据库。评估主动远端后和随访期间磨牙位置变化的研究包括在内。使用ROBINS-I工具对非随机研究进行数据提取和偏倚风险评估。进行了一项荟萃分析，以量化垂直和矢状维度的复发。结果：从最初确定的2594项研究中，6项研究（均为非随机对照试验）被纳入系统评价，5项研究被纳入meta分析。所有研究均显示中等偏倚风险。矢状面平均线性复发1.58±0.84 mm(35%)，角复发2.98±2.07°（53%）。垂直方向复发1.44±1.26 mm。荟萃分析显示，长期矢状面磨牙位置净远端为2.77 mm (95% CI 2.33-3.20)，远端为1°（95% CI -2.37°至4.38°），垂直平面为-0.65 mm （95% CI -2.27至0.96）。I2在60% ~ 98%之间，异质性较高。复发是最小的切，骨骼和软组织参数。结论：拔除磨牙后的长期评估证实了矢状面和垂直方向复发的确凿证据。垂直方向复发率高于矢状方向。在横向上评估复发的文献很少。因此，本系统综述提倡术后长期保留的必要性。

{"title":"Long term stability of molar position postdistalization: A systematic review and meta-analysis","authors":"Kutraaleeshwaran Velmurugan, Venkateswaran Ananthanarayanan, Sridevi Padmanabhan, Vignesh Kailasam","doi":"10.1016/j.ejwf.2025.07.734","DOIUrl":"10.1016/j.ejwf.2025.07.734","url":null,"abstract":"<div><h3>Background</h3><div>Molar distalization has generated additional attention with the advent of skeletal anchorage devices. However, relapse after distalization is one of the challenges and the quantum of relapse following distalization, especially over the long-term remains controversial. This systematic review and meta-analysis aim to evaluate the long-term vertical and sagittal stability of molars postdistalization.</div></div><div><h3>Methods</h3><div>A systematic search was conducted in electronic databases (PubMed, Scopus, Ovid, Embase, Cochrane and LILACS) up to April 11, 2025. Studies assessing molar position changes after active distalization and during follow-up periods were included. Data extraction and risk of bias assessments were performed using the ROBINS-I tool for non-randomized studies. A meta-analysis was conducted to quantify relapse in vertical and sagittal dimensions.</div></div><div><h3>Results</h3><div>From the 2594 studies initially identified, six studies (all nonrandomized control trials) were included for the systematic review and five studies were included for the meta-analysis. All studies showed moderate risk of bias. The mean relapse in the sagittal plane was 1.58 ± 0.84 mm (35%) linearly and 2.98 ± 2.07° (53%) angularly. In the vertical plane, it relapsed by 1.44 ± 1.26 mm. Meta-analysis showed a net distalization in the molar position in long-term sagittal plane by 2.77 mm (95% CI 2.33–3.20) and 1° distally (95% CI −2.37° to 4.38°), and in the vertical plane by −0.65 mm (95% CI −2.27 to 0.96). The I<sup>2</sup> ranged from 60% to 98% indicating high heterogeneity. The relapse was minimal for the incisal, skeletal, and soft tissues parameters.</div></div><div><h3>Conclusions</h3><div>Long-term assessment following molar distalization demonstrated conclusive evidence of relapse in the sagittal and vertical direction. There was greater relapse in the vertical than in the sagittal direction. There was minimal literature assessing the relapse in the transverse direction. Hence, this systematic review advocates the need for long-term retention postdistalization.</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Pages 355-366"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145138789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Voluntary Service, Life, and Legacy: The Emotions That Last a Lifetime and Impact Others' Lives! 志愿服务、生活和遗产：影响他人一生的情感！

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-10-18 DOI: 10.1016/j.ejwf.2025.10.002

Nikhillesh Vaiid

引用次数: 0

46 - Clinically oriented automatic three- dimensional enamel segmentation via deep learning 临床导向的牙釉质三维自动深度学习分割

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-12-09 DOI: 10.1016/j.ejwf.2025.07.050

Wenting Yu , Huifang Yang , Xianju Xie , Ning Zhang , Yuxing Bai

Background

Establishing accurate, reliable, and convenient methods for enamel segmentation and analysis is crucial for effectively planning orthodontic treatments. However, no mature, non-destructive method currently exists in clinical dentistry to quickly and accurately assess the integrity and thickness of enamel chair-side. This study aims to develop a deep learning work, 2.5D Attention U-Net, trained on small sample datasets, for the automatical and accurate segmentation of enamel in clinical settings.

Methods

We propose an automated computer-aided enamel segmentation model based on an instance segmentation network, 2.5D Attention U-Net. The model is trained using manually annotated segmented enamel data, and its performance is evaluated using the Dice metrics. A satisfactory image segmentation model is applied to generate a 3D enamel model for each tooth and to calculate the thickness value of individual enclosed 3D enamel meshes.

Results

The model achieves Dice score on the enamel segmentation task of 96.6%. This study provides an intuitive visualization of irregular enamel morphology and a quantitative analysis of three-dimensional enamel thickness variations. The results indicate that enamel is thickest at the incisal edges of anterior teeth and the cusps of posterior teeth, thinning towards the roots. For posterior teeth, the enamel is thinnest at the central fossae area, with mandibular molars having thicker enamel in the central fossae compared to maxillary molars. The average enamel thickness of maxillary incisors, canines, and premolars is greater than that of mandibular incisors, while the opposite is true for molars. The average enamel thickness gradually increases from the incisors to the molars within the same quadrant.

Conclusions

This study introduces an automatic, efficient, and accurate 2.5D Attention U-Net system to enhance precise and efficient chair-side diagnosis and treatment of enamel-related diseases in clinical settings.

背景建立准确、可靠、便捷的牙釉质分割和分析方法对于有效规划正畸治疗至关重要。然而，目前临床上还没有成熟的、无损的方法来快速准确地评估椅侧牙釉质的完整性和厚度。本研究旨在开发一个深度学习工作，2.5D注意力U-Net，在小样本数据集上训练，用于临床环境中牙釉质的自动准确分割。方法提出一种基于实例分割网络2.5D Attention U-Net的计算机辅助牙釉质自动分割模型。该模型使用人工标注的牙釉质分割数据进行训练，并使用Dice指标评估其性能。应用满意的图像分割模型生成每颗牙齿的三维牙釉质模型，并计算各个封闭的三维牙釉质网格的厚度值。结果该模型对牙釉质分割任务的Dice评分达到96.6%。本研究提供了不规则牙釉质形态的直观可视化和三维牙釉质厚度变化的定量分析。结果表明，牙釉质在前牙切缘和后牙尖处最厚，在牙根处越薄。对于后牙，牙釉质在中央窝区最薄，下颌磨牙在中央窝区的牙釉质较上颌磨牙厚。上颌门牙、犬齿和前磨牙的平均牙釉质厚度大于下颌门牙，而磨牙的平均牙釉质厚度则相反。在同一象限内，从门牙到磨牙，牙釉质的平均厚度逐渐增加。结论本研究引入了一种自动、高效、准确的2.5D注意力U-Net系统，以提高临床对牙釉质相关疾病的准确、高效的椅侧诊断和治疗。

{"title":"46 - Clinically oriented automatic three- dimensional enamel segmentation via deep learning","authors":"Wenting Yu , Huifang Yang , Xianju Xie , Ning Zhang , Yuxing Bai","doi":"10.1016/j.ejwf.2025.07.050","DOIUrl":"10.1016/j.ejwf.2025.07.050","url":null,"abstract":"<div><h3>Background</h3><div>Establishing accurate, reliable, and convenient methods for enamel segmentation and analysis is crucial for effectively planning orthodontic treatments. However, no mature, non-destructive method currently exists in clinical dentistry to quickly and accurately assess the integrity and thickness of enamel chair-side. This study aims to develop a deep learning work, 2.5D Attention U-Net, trained on small sample datasets, for the automatical and accurate segmentation of enamel in clinical settings.</div></div><div><h3>Methods</h3><div>We propose an automated computer-aided enamel segmentation model based on an instance segmentation network, 2.5D Attention U-Net. The model is trained using manually annotated segmented enamel data, and its performance is evaluated using the Dice metrics. A satisfactory image segmentation model is applied to generate a 3D enamel model for each tooth and to calculate the thickness value of individual enclosed 3D enamel meshes.</div></div><div><h3>Results</h3><div>The model achieves Dice score on the enamel segmentation task of 96.6%. This study provides an intuitive visualization of irregular enamel morphology and a quantitative analysis of three-dimensional enamel thickness variations. The results indicate that enamel is thickest at the incisal edges of anterior teeth and the cusps of posterior teeth, thinning towards the roots. For posterior teeth, the enamel is thinnest at the central fossae area, with mandibular molars having thicker enamel in the central fossae compared to maxillary molars. The average enamel thickness of maxillary incisors, canines, and premolars is greater than that of mandibular incisors, while the opposite is true for molars. The average enamel thickness gradually increases from the incisors to the molars within the same quadrant.</div></div><div><h3>Conclusions</h3><div>This study introduces an automatic, efficient, and accurate 2.5D Attention U-Net system to enhance precise and efficient chair-side diagnosis and treatment of enamel-related diseases in clinical settings.</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Page 414"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

17 - Aesthetic and social impact of anterior open bite in children and pre-adolescents 儿童和青少年前牙开咬的美学和社会影响

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-12-09 DOI: 10.1016/j.ejwf.2025.07.021

Emilly Faria , Karina Freitas , Paula Cotrin , Fabricio Valarelli , Célia Pinzan

Background

There is a lack of studies evaluating the potential psychological implications associated with anterior open bite (AOB), such as the impact on self-esteem and social acceptance, particularly across different developmental stages. Understanding the aesthetic and social perception of AOB in childhood can provide valuable insights for early intervention and improve orthodontic care.

Objective

The aim of this cross-sectional study was to evaluate the aesthetic and social impaction of anterior open bite (AOB) among children aged 4 to 12 years.

Materials and Methods

The sample included 272 evaluators (147 girls and 125 boys), divided into three age groups: 4–6, 7–9, and 10–12 years. Images of children's faces from different age groups were created using artificial intelligence, and their smiles were digitally edited using Adobe Photoshop to show AOB and normal overbite. Participants of the study completed a questionnaire to evaluate their aesthetic perception of the images using a Escala numérica numerical rating scale ranging from 0 to 10. Social perception was evaluated with objective questions. The tests t-independent and chi-cuadrado assessed the aesthetic and social impactions, respectively.

Results

The photographs of children with normal overbite showed a statistically significant more positive aesthetic perception, for all evaluated age groups (p<0.001). Regarding social perception, in the 4–6 age group, only the "Beautiful Smile" characteristic showed a statistically significant difference (p=0.002). In the 7–9 and 10–12 age groups, the characteristics "Beautiful Smile" and "Beautiful Teeth" demonstrated statistically significant differences between the images (p<0.001).

Conclusion

Anterior open bite negatively impacts the aesthetic perception of children between 4 and 12 years of age.

目前缺乏评估前牙开咬（AOB）潜在心理影响的研究，例如对自尊和社会接受的影响，特别是在不同的发育阶段。了解儿童AOB的审美和社会感知，可以为早期干预和改善正畸护理提供有价值的见解。目的本研究旨在评估4 ~ 12岁儿童前牙开咬（AOB）的美学和社会影响。材料与方法272名评价者（女生147名，男生125名），分为4 ~ 6岁、7 ~ 9岁和10 ~ 12岁3个年龄组。研究人员使用人工智能创建了不同年龄段儿童的面部图像，并使用Adobe Photoshop对他们的微笑进行了数字编辑，以显示AOB和正常的复咬合。该研究的参与者完成了一份调查问卷，以评估他们对这些图像的审美感知，使用的是从0到10的Escala numsamica数字评分量表。用客观问题评估社会知觉。t独立测试和chi-cuadrado测试分别评估审美和社会影响。结果在所有被评估的年龄组中，正常覆咬合儿童的照片显示出具有统计学意义的更积极的审美知觉（p<0.001）。在社会知觉方面，在4-6岁年龄组中，只有“美丽的微笑”特征有统计学差异（p=0.002）。在7-9岁和10-12岁年龄组中，“美丽的微笑”和“美丽的牙齿”特征在图像之间表现出统计学上的显著差异（p<0.001）。结论前牙开咬对4 ~ 12岁儿童的审美有负面影响。

{"title":"17 - Aesthetic and social impact of anterior open bite in children and pre-adolescents","authors":"Emilly Faria , Karina Freitas , Paula Cotrin , Fabricio Valarelli , Célia Pinzan","doi":"10.1016/j.ejwf.2025.07.021","DOIUrl":"10.1016/j.ejwf.2025.07.021","url":null,"abstract":"<div><h3>Background</h3><div>There is a lack of studies evaluating the potential psychological implications associated with anterior open bite (AOB), such as the impact on self-esteem and social acceptance, particularly across different developmental stages. Understanding the aesthetic and social perception of AOB in childhood can provide valuable insights for early intervention and improve orthodontic care.</div></div><div><h3>Objective</h3><div>The aim of this cross-sectional study was to evaluate the aesthetic and social impaction of anterior open bite (AOB) among children aged 4 to 12 years.</div></div><div><h3>Materials and Methods</h3><div>The sample included 272 evaluators (147 girls and 125 boys), divided into three age groups: 4–6, 7–9, and 10–12 years. Images of children's faces from different age groups were created using artificial intelligence, and their smiles were digitally edited using Adobe Photoshop to show AOB and normal overbite. Participants of the study completed a questionnaire to evaluate their aesthetic perception of the images using a Escala numérica numerical rating scale ranging from 0 to 10. Social perception was evaluated with objective questions. The tests t-independent and chi-cuadrado assessed the aesthetic and social impactions, respectively.</div></div><div><h3>Results</h3><div>The photographs of children with normal overbite showed a statistically significant more positive aesthetic perception, for all evaluated age groups (p<0.001). Regarding social perception, in the 4–6 age group, only the \"Beautiful Smile\" characteristic showed a statistically significant difference (p=0.002). In the 7–9 and 10–12 age groups, the characteristics \"Beautiful Smile\" and \"Beautiful Teeth\" demonstrated statistically significant differences between the images (p<0.001).</div></div><div><h3>Conclusion</h3><div>Anterior open bite negatively impacts the aesthetic perception of children between 4 and 12 years of age.</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Page 403"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705404","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

58 - Comparison of treatment effectiveness to time ratio between comprehensive and two different hybrid aligner treatment plan simulations 综合和两种不同混合矫正器治疗方案模拟的治疗效果与时间比的比较

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-12-09 DOI: 10.1016/j.ejwf.2025.07.062

Venkata Yudhistar Palla , Sandhya Reddy Rajula , Vdy Attelli , Ojass kumar , Vasavi Lakshmi Prasanna Mukku , Sravani Yerraginnela

Aligner treatment has transformed orthodontics by offering a modern, patient-friendly alternative to traditional Fixed Orthodontics. While aligners provide aesthetic and comfortable treatment, they can be costlier and take longer for moderate to severe cases. This in-vitro study compares the treatment-effectiveness-to-time ratio (TTR) of three orthodontic treatment plans simulated using the same software: 1.Aligner therapy 2.Hybrid Aligner 1: Initial fixed orthodontic treatment simulation (upper and lower) from Stage 1, followed by aligner therapy. 3.Hybrid Aligner 2: Aligner therapy for the maxilla and fixed orthodontic simulation for the mandible from start of treatment. The study analyzed 20 STL files with a Peer Assessment Rating (PAR) index of ≥30 and an Index of Orthodontic Treatment Need (IOTN) score of ≥3. Simulations were conducted using Archform software, and their TTR was assessed. Results indicate that Hybrid Aligner 1 achieved the best TTR, demonstrating that combining fixed appliances and aligners provides an efficient, aesthetic, and cost-effective approach, especially for high-PAR cases. These findings highlight the importance of selecting appropriate treatment strategies based on malocclusion severity, patient preferences, and resource availability. Moreover, the study suggests that an algorithm based on TTR could aid in AI-driven treatment planning, reinforcing the orthodontist’s role in managing complex cases with precision using hybrid aligners.

矫正器治疗通过提供一个现代的，病人友好的替代传统固定正畸改变了正畸。虽然矫正器提供美观和舒适的治疗，但对于中度到重度病例，它们可能更昂贵，需要更长的时间。本体外研究比较了使用同一软件模拟的三种正畸治疗方案的治疗效果-时间比（TTR）：矫正器疗法混合矫正器1：第1阶段的初始固定正畸治疗模拟（上、下），随后进行矫正器治疗。3.混合矫正器2：上颌矫正器治疗和下颌骨固定正畸模拟治疗。本研究分析了20份同行评估评分（PAR）指数≥30，正畸治疗需求指数（IOTN）评分≥3的STL文件。采用Archform软件进行模拟，并对其TTR进行评估。结果表明，Hybrid Aligner 1达到了最佳的TTR，表明将固定矫治器和矫正器结合使用提供了一种高效、美观、经济的方法，特别是对于高par病例。这些发现强调了根据错颌严重程度、患者偏好和资源可用性选择适当治疗策略的重要性。此外，该研究表明，基于TTR的算法可以帮助人工智能驱动的治疗计划，加强正畸医生在使用混合矫正器精确管理复杂病例中的作用。

{"title":"58 - Comparison of treatment effectiveness to time ratio between comprehensive and two different hybrid aligner treatment plan simulations","authors":"Venkata Yudhistar Palla , Sandhya Reddy Rajula , Vdy Attelli , Ojass kumar , Vasavi Lakshmi Prasanna Mukku , Sravani Yerraginnela","doi":"10.1016/j.ejwf.2025.07.062","DOIUrl":"10.1016/j.ejwf.2025.07.062","url":null,"abstract":"<div><div>Aligner treatment has transformed orthodontics by offering a modern, patient-friendly alternative to traditional Fixed Orthodontics. While aligners provide aesthetic and comfortable treatment, they can be costlier and take longer for moderate to severe cases. This in-vitro study compares the treatment-effectiveness-to-time ratio (TTR) of three orthodontic treatment plans simulated using the same software: 1.Aligner therapy 2.Hybrid Aligner 1: Initial fixed orthodontic treatment simulation (upper and lower) from Stage 1, followed by aligner therapy. 3.Hybrid Aligner 2: Aligner therapy for the maxilla and fixed orthodontic simulation for the mandible from start of treatment. The study analyzed 20 STL files with a Peer Assessment Rating (PAR) index of ≥30 and an Index of Orthodontic Treatment Need (IOTN) score of ≥3. Simulations were conducted using Archform software, and their TTR was assessed. Results indicate that Hybrid Aligner 1 achieved the best TTR, demonstrating that combining fixed appliances and aligners provides an efficient, aesthetic, and cost-effective approach, especially for high-PAR cases. These findings highlight the importance of selecting appropriate treatment strategies based on malocclusion severity, patient preferences, and resource availability. Moreover, the study suggests that an algorithm based on TTR could aid in AI-driven treatment planning, reinforcing the orthodontist’s role in managing complex cases with precision using hybrid aligners.</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Page 418"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705379","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

36 - Automated analysis of facial symmetry using artificial intelligence 36 -使用人工智能自动分析面部对称性

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-12-09 DOI: 10.1016/j.ejwf.2025.07.040

Mohammad Behnaz , Shahab Kavousinejad , Kazem Dalaei , Mohammadreza Safavi , Hoori Mirmohammadsadeghi , Asghar Ebadifar

Precise identification of facial asymmetry is essential in orthodontics and craniofacial diagnostics. This study proposes an automated pipeline leveraging a Siamese Neural Network (SNN) in conjunction with Exploratory Data Analysis (EDA) to detect and quantify facial asymmetry from frontal images. A dataset of 1,200 facial photographs—annotated by expert orthodontists—was used to train and evaluate the system. Facial landmarks were extracted using the MediaPipe model, followed by midline alignment based on ocular centers. EDA methods quantified asymmetry by computing normalized distances from facial key points to the midline in both horizontal and vertical directions. These metrics were further analyzed using statistical techniques such as kernel density estimation and Wasserstein distance. The SNN model, trained on composite mirrored halves of faces, demonstrated strong diagnostic performance, achieving an accuracy of 97% and an AUC of 0.98. Agreement between the model's predictions and expert evaluations was high (Cohen’s Kappa = 0.84). The framework enables objective, non-invasive, and real-time assessment of asymmetry, with potential applications in clinical and research settings.

准确识别面部不对称在正畸和颅面诊断中是必不可少的。本研究提出了一种利用连体神经网络（SNN）与探索性数据分析（EDA）相结合的自动化管道，以检测和量化正面图像中的面部不对称。一个由1200张面部照片组成的数据集——由专家正畸医生注释——被用来训练和评估该系统。使用MediaPipe模型提取面部标志，然后根据眼中心进行中线对齐。EDA方法通过计算从面部关键点到中线在水平和垂直方向上的归一化距离来量化不对称性。使用核密度估计和Wasserstein距离等统计技术进一步分析这些指标。SNN模型在人脸的复合镜像半部分上训练，显示出很强的诊断性能，准确率达到97%，AUC为0.98。模型预测与专家评价之间的一致性很高（Cohen’s Kappa = 0.84）。该框架能够实现客观、非侵入性和实时的不对称评估，在临床和研究环境中具有潜在的应用前景。

{"title":"36 - Automated analysis of facial symmetry using artificial intelligence","authors":"Mohammad Behnaz , Shahab Kavousinejad , Kazem Dalaei , Mohammadreza Safavi , Hoori Mirmohammadsadeghi , Asghar Ebadifar","doi":"10.1016/j.ejwf.2025.07.040","DOIUrl":"10.1016/j.ejwf.2025.07.040","url":null,"abstract":"<div><div>Precise identification of facial asymmetry is essential in orthodontics and craniofacial diagnostics. This study proposes an automated pipeline leveraging a Siamese Neural Network (SNN) in conjunction with Exploratory Data Analysis (EDA) to detect and quantify facial asymmetry from frontal images. A dataset of 1,200 facial photographs—annotated by expert orthodontists—was used to train and evaluate the system. Facial landmarks were extracted using the MediaPipe model, followed by midline alignment based on ocular centers. EDA methods quantified asymmetry by computing normalized distances from facial key points to the midline in both horizontal and vertical directions. These metrics were further analyzed using statistical techniques such as kernel density estimation and Wasserstein distance. The SNN model, trained on composite mirrored halves of faces, demonstrated strong diagnostic performance, achieving an accuracy of 97% and an AUC of 0.98. Agreement between the model's predictions and expert evaluations was high (Cohen’s Kappa = 0.84). The framework enables objective, non-invasive, and real-time assessment of asymmetry, with potential applications in clinical and research settings.</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Page 410"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705590","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

39 - Biological responses to corticopuncture-assisted palatal expansion in mature adult rats 39 -皮质穿刺辅助腭扩张在成年大鼠中的生物学反应

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-12-09 DOI: 10.1016/j.ejwf.2025.07.043

Romney Guimaraes , Agnaldo Silva Garcez Segundo , Kee-Joon Lee , Selly Sayuri Suzuki

Mini-implant-assisted rapid palatal expansion (MARPE) is an age-dependent technique, primarily due to the increased interdigitation of the midpalatal suture with aging, which leads to greater resistance to expansion. This study aimed to assess the biological response of the corticopuncture method, a technique designed to weaken the midpalatal suture, during maxillary expansion in mature adult rats. A total of 42 rats were included in the study and divided into three groups: a control group (CN) - no expansion or corticopuncture (n = 6), a rapid maxillary expansion group (ERM), and a rapid maxillary expansion + corticopuncture group (ERM+CP). The animals were euthanized at baseline (day 0 for the CN group) and at 3, 7, and 14 days post-expansion (ERM and ERM+CP groups). Microtomographic, histological, and immunohistochemical analyses were conducted to assess the biological responses. Both ERM and ERM+CP groups exhibited significant differences compared to the CN group across all measurements analyzed. Notably, ERM+CP group demonstrated greater trabecular separation than ERM group at days 7 and 14 (p<0.05). In linear measurements, ERM+CP group showed significantly greater expansion compared to the ERM group (p<0.05). On the contrary, bone volume relative to total volume and bone density were lower in the ERM+CP group than in the ERM group (p<0.05). Histological and immunohistochemical analyses revealed increased osteoclastic activity in the ERM+CP group, particularly on day 3. These findings suggest that maxillary expansion assisted by corticopuncture enhances suture separation compared to expansion alone, likely due to the biological response induced by the procedure

微型种植体辅助腭快速扩张（MARPE）是一种年龄依赖性技术，主要是因为随着年龄的增长，中腭缝合线的指间性增加，从而导致更大的扩张阻力。本研究旨在评估皮质穿刺法（一种旨在削弱中腭缝合线的技术）在成年大鼠上颌扩张期间的生物学反应。将42只大鼠分为3组：对照组(CN) -无扩张或皮质穿刺（n = 6）、上颌快速扩张组（ERM）和上颌快速扩张 + 皮质穿刺组（ERM+CP）。在基线（CN组为第0天）和扩展后3、7和14天（ERM和ERM+CP组）对动物实施安乐死。通过显微层析、组织学和免疫组织化学分析来评估生物反应。与CN组相比，ERM组和ERM+CP组在所有测量结果中都表现出显著差异。值得注意的是，与ERM组相比，ERM+CP组在第7天和第14天出现了更大的小梁分离（p < 0.05）。在线性测量中，ERM+CP组的扩张明显大于ERM组（p<0.05）。相反，ERM+CP组骨体积相对于总体积和骨密度低于ERM组（p < 0.05）。组织学和免疫组织化学分析显示，ERM+CP组的破骨细胞活性增加，特别是在第3天。这些研究结果表明，与单纯扩张相比，皮质穿刺辅助上颌扩张可以增强缝线分离，这可能是由于手术引起的生物反应

{"title":"39 - Biological responses to corticopuncture-assisted palatal expansion in mature adult rats","authors":"Romney Guimaraes , Agnaldo Silva Garcez Segundo , Kee-Joon Lee , Selly Sayuri Suzuki","doi":"10.1016/j.ejwf.2025.07.043","DOIUrl":"10.1016/j.ejwf.2025.07.043","url":null,"abstract":"<div><div>Mini-implant-assisted rapid palatal expansion (MARPE) is an age-dependent technique, primarily due to the increased interdigitation of the midpalatal suture with aging, which leads to greater resistance to expansion. This study aimed to assess the biological response of the corticopuncture method, a technique designed to weaken the midpalatal suture, during maxillary expansion in mature adult rats. A total of 42 rats were included in the study and divided into three groups: a control group (CN) - no expansion or corticopuncture (n = 6), a rapid maxillary expansion group (ERM), and a rapid maxillary expansion + corticopuncture group (ERM+CP). The animals were euthanized at baseline (day 0 for the CN group) and at 3, 7, and 14 days post-expansion (ERM and ERM+CP groups). Microtomographic, histological, and immunohistochemical analyses were conducted to assess the biological responses. Both ERM and ERM+CP groups exhibited significant differences compared to the CN group across all measurements analyzed. Notably, ERM+CP group demonstrated greater trabecular separation than ERM group at days 7 and 14 (p<0.05). In linear measurements, ERM+CP group showed significantly greater expansion compared to the ERM group (p<0.05). On the contrary, bone volume relative to total volume and bone density were lower in the ERM+CP group than in the ERM group (p<0.05). Histological and immunohistochemical analyses revealed increased osteoclastic activity in the ERM+CP group, particularly on day 3. These findings suggest that maxillary expansion assisted by corticopuncture enhances suture separation compared to expansion alone, likely due to the biological response induced by the procedure</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Page 411"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705593","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

26 - Artificial intelligence for the automated detection of dental malocclusions from clinical images 用于从临床图像中自动检测牙合错误的人工智能

IF 3.2 Q1 DENTISTRY, ORAL SURGERY & MEDICINE

Journal of the World Federation of Orthodontists

Pub Date : 2025-12-01 Epub Date: 2025-12-09 DOI: 10.1016/j.ejwf.2025.07.030

Babak Sayahpour , Ani Nebiaj , Bernd Freisleben , Markus Muehling

Background: Accurate identification of dental malocclusions is essential for orthodontic care yet remains labour-intensive and examiner-dependent.

Objective: To develop and validate a deep-learning model for automated detection of malocclusions on routine clinical photographs.

Materials and Methods: A dataset of 2,077 anonymised, high-resolution intra-oral images (frontal occlusion, right/left buccal, maxillary, mandibular) was extracted from the OnyxCeph^Ⓡ database of Goethe University Frankfurt. Images were labelled for malocclusion traits according to Index of Orthodontic Treatment Need criteria and augmented by rotation, reflection, scaling, brightness variation and noise. An advanced YOLOv11 object-detection network was trained with an 80/10/10 split and evaluated by mean average precision at 0.5 intersection-over-union (mAP-50).

Results: The model achieved an mAP-50 = 0.854, corresponding to 85 % agreement with specialist diagnoses. Average inference time was 0.05 s on a consumer-grade GPU, enabling real-time chairside use.

Conclusions: This pilot study shows that deep-learning object detection can reliably identify common malocclusions from photographic records and could streamline orthodontic diagnostics. Future work should assess clinical impact using larger multicentre datasets and iterative network optimisation.

背景：准确识别牙齿错颌是必不可少的正畸护理，但仍然是劳动密集型和检查员依赖。目的：建立并验证一种用于临床常规照片错颌自动检测的深度学习模型。材料和方法：从法兰克福歌德大学的OnyxCephⓇ数据库中提取2,077张匿名的高分辨率口腔内图像（额颌、右/左颊、上颌、下颌）。根据正畸治疗需求指数（Index of Orthodontic Treatment Need）标准标记图像的错颌特征，并通过旋转、反射、缩放、亮度变化和噪声进行增强。采用80/10/10分割训练先进的YOLOv11目标检测网络，并以0.5相交-超并（mAP-50）的平均精度进行评估。结果：模型达到了mAP-50 = 0.854，与专家诊断的符合率达到85%。在消费级GPU上，平均推理时间为0.05秒，支持实时椅边使用。结论：本初步研究表明，深度学习对象检测可以可靠地从照片记录中识别常见的咬合错误，并可以简化正畸诊断。未来的工作应该使用更大的多中心数据集和迭代网络优化来评估临床影响。

{"title":"26 - Artificial intelligence for the automated detection of dental malocclusions from clinical images","authors":"Babak Sayahpour , Ani Nebiaj , Bernd Freisleben , Markus Muehling","doi":"10.1016/j.ejwf.2025.07.030","DOIUrl":"10.1016/j.ejwf.2025.07.030","url":null,"abstract":"<div><div><strong>Background:</strong> Accurate identification of dental malocclusions is essential for orthodontic care yet remains labour-intensive and examiner-dependent.</div><div><strong>Objective:</strong> To develop and validate a deep-learning model for automated detection of malocclusions on routine clinical photographs.</div><div><strong>Materials and Methods:</strong> A dataset of 2,077 anonymised, high-resolution intra-oral images (frontal occlusion, right/left buccal, maxillary, mandibular) was extracted from the OnyxCeph<sup>Ⓡ</sup> database of Goethe University Frankfurt. Images were labelled for malocclusion traits according to Index of Orthodontic Treatment Need criteria and augmented by rotation, reflection, scaling, brightness variation and noise. An advanced YOLOv11 object-detection network was trained with an 80/10/10 split and evaluated by mean average precision at 0.5 intersection-over-union (mAP-50).</div><div><strong>Results:</strong> The model achieved an mAP-50 = 0.854, corresponding to 85 % agreement with specialist diagnoses. Average inference time was 0.05 s on a consumer-grade GPU, enabling real-time chairside use.</div><div><strong>Conclusions:</strong> This pilot study shows that deep-learning object detection can reliably identify common malocclusions from photographic records and could streamline orthodontic diagnostics. Future work should assess clinical impact using larger multicentre datasets and iterative network optimisation.</div></div>","PeriodicalId":43456,"journal":{"name":"Journal of the World Federation of Orthodontists","volume":"14 6","pages":"Page 406"},"PeriodicalIF":3.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705657","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0