基于机器学习的颌骨特征点还原法辅助跨中线颌骨缺损重建

doi:10.19438/j.cjoms.2020.04.007

中国口腔颌面外科杂志 ›› 2020, Vol. 18 ›› Issue (4): 323-327.doi: 10.19438/j.cjoms.2020.04.007

基于机器学习的颌骨特征点还原法辅助跨中线颌骨缺损重建

周子疌¹, 朱向阳², 韩婧¹, 刘剑楠^1,*, 张陈平^1,*

1.上海交通大学医学院附属第九人民医院·口腔医学院口腔颌面-头颈肿瘤科,国家口腔疾病临床医学研究中心,上海市口腔医学重点实验室,上海市口腔医学研究所,上海 200011;
2.上海交通大学图像通信与网络工程学院,上海 200240

收稿日期:2020-02-05 出版日期:2020-07-20 发布日期:2020-09-10
通讯作者: 张陈平, E-mail:zhang.chenping@hotmail.com;刘剑楠,E-mail:laurence_ljn@163.com。*共同通信作者
作者简介:周子疌（1994-）,男,硕士研究生,E-mail: plenilunezzj@163.com
基金资助:
国家重点研发计划（2018YFC0807303）

Landmarks restore,a machine learning algorithm-aided surgical planning for cross-midline maxillo-mandibular defect

ZHOU Zi-jie¹, ZHU Xiang-yang², HAN Jing¹, LIU Jian-nan¹, ZHANG Chen-ping¹

1. Department of Oromaxillofacial Head and Neck Oncology, Shanghai Ninth People's Hospital, College of Stomatology, Shanghai Jiao Tong University School of Medicine; National Clinical Research Center for Oral Diseases; Shanghai Key Laboratory of Stomatology & Shanghai Research Institute of Stomatology. Shanghai 200011;
2. Institute of Image Communication and Network Engineering, Shanghai Jiao Tong University. Shanghai 200240, China

Received:2020-02-05 Online:2020-07-20 Published:2020-09-10

摘要/Abstract

摘要： 目的: 通过测量正常汉族人群颌骨关键特征点数据,计算并分析上、下颌骨间各特征点的内在联系,为跨中线颌骨缺损的个体化重建提供参考。方法: 收集111例华东地区正常成年人颌骨CT数据（Dicom格式）,其中男43例、女68例,平均年龄24.3岁。应用手术规划系统（Proplan CMF 3.0）进行上、下颌骨分割,同时标记并收集16个颌骨特征点空间坐标（上颌骨9个、下颌骨7个）。借助MATLAB数学软件,对颌骨特征点坐标进行统计分析;并通过机器学习算法,研究特征点坐标间的相关性。采用SPSS 22.0软件包对男女颌骨外形数据进行t 检验,分析性别差异。结果: 不同性别的颌骨特征点的线性参数存在显著差异（P<0.05）。除∠b₁ab₂（男性136.06°,女性132.18°,P<0.05）外,角度变量间无显著性别差异。基于矢量相似度匹配模型完成机器学习算法开发,并将该算法用于1例下颌骨跨中线缺损患者的术前设计。结论: 机器学习的颌骨特征点还原法可在复杂颌骨重建过程中提供精确的个体化方案,解决跨中线大范围颌骨缺损病例重建仅凭经验、无参照可依的临床难题。

关键词: 颌骨重建, 解剖标志点, 机器学习, 术前规划

Abstract: PURPOSE: The distance and angles among characteristic points of normal Chinese jaws were measured and analyzed by an algorithm to explore the morphological correlation between maxilla and mandible,and to guide a personalized preoperative design for cross-midline defects. METHODS: A total of 111 normal maxillary CT data (Dicom format) were collected and delivered to surgical planning software(Proplan CMF 3.0),which included 43 males and 68 females. Each case had 16 key points to mark. The data were analyzed using Mathematical software(MATLAB). The data from male and female jaw bones were analyzed using SPSS 22.0 software package to analyze the difference between genders. RESULTS: There was significant difference (P<0.05) in the size of jaw bones between genders, but no significant difference in angles except for ∠b₁ab₂(males 136.06°,females 132.18°,P<0.05 ). A machine learning algorithm was programmed based on vector matching model, and it was clinically applied to a patient with cross-midline mandibular defect. CONCLUSIONS: With the benefit of algorithm-aided preoperative design, a more convenient way can be achieved to tailor a similar maxilla or mandible for patients, especially in cases with cross-midline defects.

Key words: Maxillo-mandibular reconstruction, Anatomic landmarks, Machine learning, Preoperative planning

中图分类号:

R782.2

周子疌, 朱向阳, 韩婧, 刘剑楠, 张陈平. 基于机器学习的颌骨特征点还原法辅助跨中线颌骨缺损重建[J]. 中国口腔颌面外科杂志, 2020, 18(4): 323-327.

ZHOU Zi-jie, ZHU Xiang-yang, HAN Jing, LIU Jian-nan, ZHANG Chen-ping. Landmarks restore,a machine learning algorithm-aided surgical planning for cross-midline maxillo-mandibular defect[J]. China J Oral Maxillofac Surg, 2020, 18(4): 323-327.

参考文献

[1] Peng X, Mao C, Yu GY, et al.Maxillary reconstruction with the free fibula flap[J]. Plast Reconstr Surg, 2005, 115(6): 1562-1569.
[2] Bak M, Jacobson AS, Buchbinder D, et al.Contemporary reconstruction of the mandible[J]. Oral Oncol, 2010, 46(2): 71-76.
[3] Brown JS, Shaw RJ.Reconstruction of the maxilla and midface: introducing a new classification[J]. Lancet Oncol, 2010,11(10):1001-1008.
[4] Brown JS, Barry C, Ho M, et al.A new classification for mandibular defects after oncological resection[J]. Lancet Oncol, 2016,17(1): e23-30.
[5] Ong HS, Yu M, Liu JN, et al.Modified maxillary reconstruction technique for Brown's class Ⅲ defects by subdividing a rhomboid shaped iliac crest into 2 subunits[J]. Head Neck, 2018, 40(8): 1824-1833.
[6] Xu LQ, Zhang CP, Poh EH, et al.A novel fibula osteotomy guide for mandibular reconstruction[J]. Plast Reconstr Surg, 2012, 129(5): 861e-863e.
[7] Bähr W.Blood supply of small fibula segments: An experimental study on human cadavers[J]. J Craniomaxillofac Surg, 1998, 26(3): 148-152.
[8] 张润, 王永滨. 机器学习及其算法和发展研究[J]. 中国传媒大学学报(自然科学版),2016,23(2):10-24.
[9] Rajkomar A, Dean J, Kohane I, et al.Machine Learning in Medicine[J]. N Engl J Med, 2019, 380(14): 1347-1358.
[10] Komura D, Ishikawa S.Machine learning methods for histopathological image analysis[J]. Comput Struct Biotechnol J, 2018, 16: 34-42.
[11] Campanella G, Hanna MG, Geneslaw L, et al.Clinical-grade computational pathology using weakly supervised deep learning on whole slide images[J]. Nat Med, 2019, 25(8): 1301-1309.
[12] Zhu X, Han J, Zhang S, et al. Restoring skeletal marker points for severe maxillary and mandibular jaw defects using a linear regression approach [J]. J Oral Maxillofac Surg, 2019, 77(3): 664.e1-664.e16.
[13] Winder J, Bibb R.Medical rapid prototyping technologies: state of the artand current limitations for application in oral and maxillofacial surgery[J]. J Oral Maxillofac Surg, 2005, 63(7): 1006-1015.
[14] 王绍义, 蒋欣泉, 张志愿. 快速成形术在下颌骨缺损修复中的应用进展[J]. 中国口腔颌面外科杂志, 2007, 5(5): 381-385.
[15] Lee JW, Fang JJ, Chang LR, et al.Mandibular defect reconstruction with the help of mirror imaging coupled with laser stereolithographic modeling technique[J]. Formos Med Assoc, 2007, 106(3): 244-250.
[16] 张海钟, 步荣发, 柳春明, 等. 中国北方美貌女性颅面骨三维测量数据库的建立[J]. 中华整形外科杂志, 2007, 23(2): 130-134.
[17] 于淼, 王慧珊, 韩婧, 等. 应用两亚单元塑形血管化髂骨-腹内斜肌瓣行全上颌骨重建[J]. 中国口腔颌面外科杂志, 2016, 14(5): 430-434.

基于机器学习的颌骨特征点还原法辅助跨中线颌骨缺损重建

Landmarks restore,a machine learning algorithm-aided surgical planning for cross-midline maxillo-mandibular defect

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 10

编辑推荐

Metrics

本文评价

[1]	杨宗燕, 康一帆, 吕晓鸣, 张雷, 单小峰, 蔡志刚. 血管化髂骨瓣重建下颌骨有牙区缺损时髂骨嵴不同朝向的优缺点比较[J]. 中国口腔颌面外科杂志, 2021, 19(5): 434-438.
[2]	杨鑫, 吕明明, 李军, 孙坚. 虚拟手术计划在经口内入路行下颌骨切除及精确重建中的应用：附9例报道[J]. 中国口腔颌面外科杂志, 2021, 19(5): 469-473.
[3]	文振宇, 郭雯瑾, 冯爱民, 李龙德, 文世生. 基于机器学习的智牙拔除后疗效预测及分析[J]. 中国口腔颌面外科杂志, 2021, 19(2): 156-162.
[4]	王明一, 傅振, 张陈平. 血管化腓骨重建下颌骨后用腓骨残端游离移植及牙种植的可行性研究[J]. 中国口腔颌面外科杂志, 2020, 18(4): 338-342.
[5]	孙坚，沈毅，吕明明，李锋，王慧珊，白石柱，王良，杨鑫，李军. 经口内入路切除下颌骨良性肿瘤及同期血管化腓骨肌瓣精确重建下颌骨[J]. 中国口腔颌面外科杂志, 2018, 16(3): 193-198.
[6]	王明一, 王慧珊, 杨劼, 刘剑楠, 曲行舟, 张陈平. 咬合引导的颌骨重建精确度分析及功能评价[J]. 中国口腔颌面外科杂志, 2018, 16(2): 162-166.
[7]	于淼, 王慧珊, 韩婧, 胡龙威, 刘剑楠, 王洋, 张陈平, 徐立群. 应用两亚单元塑形血管化髂骨-腹内斜肌瓣行全上颌骨重建[J]. 中国口腔颌面外科杂志, 2016, 14(5): 430-434.
[8]	王宏伟，马春跃，贺光，徐立群，张陈平，秦兴军. 家族性巨大型牙骨质瘤患者下颌骨重建术式的选择:2例报告及文献复习[J]. 中国口腔颌面外科杂志, 2016, 14(1): 89-92.
[9]	周苗, 海德克·特海登, 朴正国. 正颌手术和间隔式骨移植重建严重萎缩颌骨的效果评价[J]. 中国口腔颌面外科杂志, 2015, 13(2): 140-146.
[10]	黄栋, 何冬梅, 杨驰, 陈敏洁, 袁建兵, 白果, 陈琢之. 数字化导板在下颌骨良性病变截骨并保存下牙槽神经血管束中的应用：附1例报告[J]. 中国口腔颌面外科杂志, 2014, 12(1): 45-50.