法医学杂志

• 论著 •    

基于机器学习和颅缝CT-MPR技术的北方成人年龄推断

魏璇1(), 陈雨珊1, 丁杰2, 宋长兴2, 王俊静2, 彭钊3, 邓振华1, 伊旭2(), 范飞1()   

  1. 1.四川大学华西基础医学与法医学院,四川 成都 610041
    2.秦皇岛市北戴河医院影像科,河北 秦皇岛 066100
    3.四川大学华西医院放射科,四川 成都 610041
  • 收稿日期:2023-12-26
  • 通讯作者: 伊旭,范飞
  • 作者简介:魏璇(2002—),女,主要从事法医临床学与法医影像学研究;E-mail:weixuanmara@163.com
  • 基金资助:
    国家自然科学基金面上项目(81971801);上海市法医学重点实验室司法部司法鉴定重点实验室开放课题(KF202209);上海市现场物证重点实验室开放课题基金资助项目(2020XCWZK04)

Age Estimation by Machine Learning and Multi-Planar Reconstruction of Cranial Sutures in Northern Chinese Han Adults

Xuan WEI1(), Yu-shan CHEN1, Jie DING2, Chang-xing SONG2, Jun-jing WANG2, Zhao PENG3, Zhen-hua DENG1, Xu YI2(), Fei FAN1()   

  1. 1.West China School of Basic Medical Sciences & Forensic Medicine, Sichuan University, Chengdu 610041, China
    2.Department of Radiology, Beidaihe Hospital of Qinhuangdao, Qinhuangdao 066100, Hebei Province, China
    3.Department of Radiology, West China Hospital, Sichuan University, Chengdu 610041, China
  • Received:2023-12-26
  • Contact: Xu YI, Fei FAN

摘要:

目的 运用CT和多平面重组(multiplanar reformation,MPR)技术获取颅缝断层图像,建立北方汉族成人特异性年龄推断模型,探讨颅缝闭合规律在中国北方人群年龄推断中的适用性。 方法 回顾性收集29~80岁健康北方汉族成人头部CT样本132例。对颅骨进行容积重组(volume reconstruction,VR)和MPR,每例样本生成160张颅缝断层图像。根据颅缝闭合分级标准对颅缝MPR图像进行评分,分别计算矢状缝、左右侧冠状缝和左右侧人字缝的平均闭合等级。以上述等级为自变量,建立北方成人年龄推断的线性回归模型和梯度提升回归、支持向量回归、决策树回归和贝叶斯岭回归4种机器学习模型,并评估各模型推断年龄的准确性。 结果 各颅缝闭合等级均与年龄呈正相关,其中矢状缝相关性最高。4种机器学习模型年龄推断的准确性均高于线性回归模型,其中支持向量回归模型的准确性最高,平均绝对误差为9.542岁。 结论 机器学习模型和颅骨CT-MPR技术可联合用于中国北方成人年龄推断,但在法医学实践中仍需与其他成人年龄推断指标联合推断年龄。

关键词: 法医人类学, 年龄推断, 机器学习, 颅缝, 计算机体层成像, 多平面重组

Abstract:

Objective To establish age estimation models of northern Han adults with linear regression and machine learning using cranial suture images obtained by multi-planar reconstruction (MPR), and to explore the applicability of cranial suture fusion rule in age estimation of northern Chinese population. Methods The head CT samples of 132 northern Han adults aged 29-80 years were retrospectively collected. After volume reconstruction (VR) and corresponding MPR, 160 cranial suture tomography images were generated for each sample. Then the MPR images of cranial sutures were scored according to the fusion grading criteria, and the mean closure grades of sagittal suture, coronal suture (both left and right) and lambdoid suture (both left and right) were calculated respectively. Finally taking the above grades as independent variables, the linear regression model and four machine learning models for age estimation: gradient boosting regression, support vector regression, decision tree regression and Bayesian ridge regression were established for northern Chinese adult’s age estimation. The accuracy of each model was evaluated. Results Each cranial suture fusion grade was positively correlated with age and the correlation of sagittal suture was the highest. All four machine learning models had higher age estimation accuracy than linear regression model. The support vector regression model had the highest accuracy among the machine learning models with a mean absolute error of 9.542 years. Conclusion The combination of MPR and machine learning model can be used for adult age prediction in northern China, but it is still necessary to combine with other adult age estimation indicators in forensic practice.

Key words: forensic anthropology, age estimation, machine learning, cranial suture, computed tomography, multiplanar reformation

中图分类号: