人工智能在颌面部骨折诊疗中的应用进展

doi:10.13591/j.cnki.kqyx.2025.05.012

摘要/Abstract

摘要：

颌面部骨折是口腔颌面外科常见多发疾病。传统的临床诊治过程易受复杂颌面部解剖结构和医生阅片及诊断经验差异等因素影响。近年来,人工智能技术的发展为颌面部骨折的精确诊断和治疗方案制定提供了新解。通过计算机视觉方法自动化影像分析,提升诊断精度与效率,并辅助制定治疗方案,展现出广泛的应用前景和应用价值。本文对国内外人工智能应用在颌面部骨折辅助诊疗中的研究工作进行回顾与综述,分析其优势与不足,并对未来研究趋势提出展望。

关键词: 人工智能, 颌面部骨折, 辅助诊断, 骨折复位

Abstract:

Maxillofacial fractures are common and frequently occurring diseases in Oral and Maxillofacial Surgery. The traditional clinical diagnosis and treatment process is easily affected by complex maxillofacial anatomy and differences in doctors’ experience in reading X-rays and making diagnoses. In recent years, artificial intelligence technology has provided new solutions for the accurate diagnosis and treatment planning of maxillofacial fractures. Automating image analysis through computer vision methods improves diagnostic accuracy and efficiency and assists in formulating treatment plans, showing broad application prospects and value. This article reviews and summarizes the research on the application of artificial intelligence in the auxiliary diagnosis and treatment of maxillofacial fractures at home and abroad, analyzes its advantages and disadvantages, and looks forward to future research trends.

Key words: artificial intelligence, maxillofacial fractures, auxiliary diagnosis, fracture reduction

中图分类号:

R782.4

黄淑慧, 朱珠, 汪云毅, 徐雨悦, 李竞, 俞刚, 张峰. 人工智能在颌面部骨折诊疗中的应用进展[J]. 口腔医学, 2025, 45(5): 386-393.

HUANG Shuhui, ZHU Zhu, WANG Yunyi, XU Yuyue, LI Jing, YU Gang, ZHANG Feng. Progress in the application of artificial intelligence in the diagnosis and treatment of maxillofacial fractures[J]. Stomatology, 2025, 45(5): 386-393.

图/表 6

图1

表1

图2

图3

表2

图4

参考文献 34

[1]	Gómez Roselló E, Quiles Granado AM, Artajona Garcia M, et al. Facial fractures: Classification and highlights for a useful report[J]. Insights Imaging, 2020, 11(1): 49.
[2]	陈春霞. 浅析机器学习的发展与应用[J]. 信息系统工程, 2017(8): 99-100.
[3]	姜国睿, 陈晖, 王姝歆. 人工智能的发展历程与研究初探[J]. 计算机时代, 2020(9): 7-10, 16.
[4]	张益, 安金刚, 朱洪平, 等. 上颌骨骨折诊疗指南(2022版)[Z]. 北京: 卫生健康委员会, 2022.
[5]	张益, 安金刚, 朱洪平, 等. 上颌骨骨折诊疗指南(2022版)[Z]. 北京: 卫生健康委员会, 2022.
[6]	Nishiyama M, Ishibashi K, Ariji Y, et al. Performance of deep learning models constructed using panoramic radiographs from two hospitals to diagnose fractures of the mandibular condyle[J]. Dento Maxillo Facial Radiology, 2021, 50(7): 20200611.
[7]	Warin K, Limprasert W, Suebnukarn S, et al. Assessment of deep convolutional neural network models for mandibular fracture detection in panoramic radiographs[J]. Int J Oral Maxillofac Surg, 2022, 51(11): 1488-1494.
[8]	Vinayahalingam S, van Nistelrooij N, van Ginneken B, et al. Detection of mandibular fractures on panoramic radiographs using deep learning[J]. Sci Rep, 2022, 12(1):19596. doi: 10.1038/s41598-022-23445-w pmid: 36379971
[9]	Nam Y, Choi Y, Kang J, et al. Diagnosis of nasal bone fractures on plain radiographs via convolutional neural networks[J]. Sci Rep, 2022, 12(1): 21510.
[10]	Seol YJ, Kim YJ, Kim YS, et al. A Study on 3D deep learning-based automatic diagnosis of nasal fractures[J]. Sensors (Basel, Switzerland), 2022, 22(2): 506.
[11]	Wang XB, Xu ZN, Tong YH, et al. Detection and classification of mandibular fracture on CT scan using deep convolutional neural network[J]. Clin Oral Investig, 2022, 26(6): 4593-4601.
[12]	Moon G, Kim S, Kim W, et al. Computer aided facial bone fracture diagnosis (CA-FBFD) system based on object detection model[J]. IEEE Access, 2022, 10: 79061-79070.
[13]	Warin K, Limprasert W, Suebnukarn S, et al. Maxillofacial fracture detection and classification in computed tomography images using convolutional neural network-based models[J]. Sci Rep, 2023, 13(1): 3434. doi: 10.1038/s41598-023-30640-w pmid: 36859660
[14]	Wang HC, Wang SC, Yan JL, et al. Artificial intelligence model trained with sparse data to detect facial and cranial bone fractures from head CT[J]. J Digit Imaging, 2023, 36(4): 1408-1418.
[15]	Zhou T, Wang HW, Du YH, et al. M3YOLOv5: Feature enhanced YOLOv5 model for mandibular fracture detection[J]. Comput Biol Med, 2024, 173: 108291.
[16]	Tian TP, Chen Y, Leow WK, et al. Computing neck-shaft angle of femur for X-ray fracture detection[M]//Lecture Notesin Computer Science. Berlin, Heidelberg: Springer Berlin Heidelberg, 2003: 82-89.
[17]	Yap DWH, Chen Y, Leow WK, et al. Detecting femur fractures by texture analysis of trabeculae[C]// Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR. August 26-26, 2004,Cambridge,UK. IEEE, 2004: 730-733.
[18]	Donnelley M, Knowles G, Hearn T. A CAD system for long-bone segmentation and fracture detection[M]//Lecture Notesin Computer Science. Berlin, Heidelberg: Springer Berlin Heidelberg, 2008: 153-162.
[19]	Bandyopadhyay O, Biswas A, Bhattacharya BB. Long-bone fracture detection in digital X-ray images based on digital-geometric techniques[J]. Comput Methods Programs Biomed, 2016, 123: 2-14.
[20]	Lim SE, Xing Y, Chen Y, et al. Detection of femur and radius fractures in X-ray images[C]. Proc 2nd Int Conf on Advances in Medical Signal and Information Processing, 2004, 249-256.
[21]	Cao Y, Wang HZ, Moradi M, et al. Fracture detection in X-ray images through stacked random forests feature fusion[C]//2015 IEEE 12th International Symposium on Biomedical Imaging (ISBI). April 16-19, 2015, Brooklyn, NY, USA.IEEE, 2015: 801-805.
[22]	Umadevi N, Geethalakshmi SN. Multiple classification system for fracture detection in human bone X-ray images[C]// 2012 Third International Conference on Computing, Communication and Networking Technologies (ICCCNT'12). July 26-28, 2012, Coimbatore, India. IEEE, 2012: 1-8.
[23]	Olczak J, Fahlberg N, Maki A, et al. Artificial intelligence for analyzing orthopedic trauma radiographs[J]. Acta Orthop, 2017, 88(6): 581-586. doi: 10.1080/17453674.2017.1344459 pmid: 28681679
[24]	Ma YL, Luo YX. Bone fracture detection through the two-stage system of Crack-Sensitive Convolutional Neural Network[J]. Inform Med Unlocked, 2021, 22: 100452.
[25]	Wang YR, Lu L, Cheng CT, et al. Weakly supervised universal fracture detection in pelvic X-rays[M]// Lecture Notesin Computer Science. Cham: Springer International Publishing, 2019: 459-467.
[26]	顾徐嘉, 孟箭, 李志萍. 虚拟重建及3D打印复合导板在经口气管插管复杂颌面部骨折复位中的应用[J]. 口腔医学研究, 2023, 39(9): 785-791. doi: 10.13701/j.cnki.kqyxyj.2023.09.005
[27]	高晓杰, 刘冰, 李冰, 等. 数字化技术结合3D打印导板在颌面部骨折手术中的应用研究[J]. 中国医学装备, 2024, 21(5): 98-101, 106.
[28]	管亚芩, 吴锦阳, 余丽雅, 等. 数字化手术设计流程在颌面部骨折整复治疗中的应用分析[J]. 组织工程与重建外科, 2023, 19(3): 289-293.
[29]	Chowdhury AS, Bhandarkar SM, Robinson RW, et al. Virtual craniofacial reconstruction from computed tomography image sequences exhibiting multiple fractures[C]// 2006 International Conference on Image Processing. October 8-11, 2006, Atlanta, GA, USA. IEEE, 2006: 1173-1176.
[30]	Luque-Luque A, Pérez-Cano FD, Jiménez-Delgado JJ. Complex fracture reduction by exact identification of the fracture zone[J]. Med Image Anal, 2021, 72: 102120.
[31]	Kronman A, Joskowicz L. Automatic bone fracture reduction by fracture contact surface identification and registration[C]// 2013 IEEE 10th International Symposium on Biomedical Imaging. April 7-11, 2013, San Francisco, CA, USA. IEEE, 2013: 246-249.
[32]	Moghari MH, Abolmaesumi P. Global registration of multiple bone fragments using statistical atlas models: Feasibility experiments[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2008, 2008: 5374-5377. doi: 10.1109/IEMBS.2008.4650429 pmid: 19163932
[33]	Han R, Uneri A, Vijayan RC, et al. Fracture reduction planning and guidance in orthopaedic trauma surgery via multi-body image registration[J]. Med Image Anal, 2021, 68: 101917.
[34]	Cleveland CN, Kelly A, DeGiovanni J, et al. Maxillofacial trauma in children: Association between age and mandibular fracture site[J]. Am J Otolaryngol, 2021, 42(2): 102874.

参考文献	数据类型	纳入人群	骨折样本量大小	数据库来源	是否公开
Nishiyama等^[6]	全景X线片	16岁以上	200例下颌骨髁突骨折患者	日本爱知学院大学和大垣市立医院	否
Warin等^[7]	全景X线片	18岁以上	855例下颌骨骨折患者	2016年—2020年某地区创伤中心	是
Vinayahalingam等^[8]	全景X线片	18岁以上	1 624例下颌骨骨折患者	德国柏林夏里特医院	否
Nam等^[9]	双侧鼻骨X线片	18岁以上	6 713例鼻骨骨折患者	2009年1月至2020年10月韩国首尔圣母医院和汝矣岛圣母医院	否
Seol等^[10]	CT轴向视图	未写明	1 185例鼻骨骨折患者	韩国嘉泉大学吉尔医疗中心医院	否
Wang等^[11]	CT轴向视图	18岁以上	686例下颌骨骨折患者	2013年1月至2020年7月北京大学口腔医院	否
Moon等^[12]	CT轴向视图	未写明	690例颌面部骨折患者	2014年—2020年韩国国立江原大学医院	否
Warin等^[13]	CT轴向视图	18岁以上	150例颌面部骨折患者	2016年—2020年某地区创伤中心	是
Wang等^[14]	CT轴向视图	5岁以上	788例颅面部骨折患者	2005年5月至2018年2月某三级创伤转诊医院	否
Zhou等^[15]	CT冠、轴、矢向视图	未写明	5 861张下颌骨骨折图像(未写明病例数)	宁夏三大医院	否

颌面部骨折自动检测
研究者	数据类型	任务对象	模型架构	研究结果
Warin K等^[7]	X线片	下颌骨	Faster R-CNN和YOLOv5	1 Faster R-CNN:Precision=0.879;Recall=0.936;F1 score=0.891 2 YOLOv5:Precision=0.861;Recall=0.922;F1 score=0.891
Vinayahalingam S等^[8]	X线片	下颌骨	结合Swin Transformer的Faster R-CNN	Precision=0.935;Recall=0.960;F1 score=0.947
Nam Y等^[9]	X线片	鼻骨	基于CNN的DL模型	Accuracy=0.833
Warin K等^[13]	轴向CT视图	颌面部	Faster R-CNN和YOLOv5	1 Faster R-CNN:Precision=0.70~0.86;Recall=0.81~0.92;F1 score=0.75~0.84 2 YOLOv5:Precision=0.00~0.84;Recall=0.00~0.92;F1 score=0.00~0.81
Moon G等^[12]	轴向CT视图	鼻骨	以YOLOX-S为基线的CA-FBFD系统	Average Precision=0.698;F1 score=0.933
Zhou T等^[15]	冠矢轴三向CT视图	下颌骨	以YOLOv5为基线的M3YOLOv5模型	Precision=0.972;Recall=0.944;F1 score=0.956
颌面部骨折自动分类
研究者	数据类型	任务对象	模型架构	研究结果
Wang XB等^[11]	轴向CT视图	下颌骨	分类模型U-Net;检测模型ResNet-50t	Dice=0.943;Accuracy=0.939~0.983
Wang HC等^[14]	轴向CT视图	颅骨及颌面部	分类模型ResUNet++;检测模型YOLOv4	检测Recall=0.887;Precision=0.945;F1 score=0.914 分割Accuracy=0.809