根据骨盆CT片特征指标进行同一认定的研究  

The study of the pelvic forensic identification on CT film

作　　者：赵峰[1] 张振[2] 丁春丽[1] 

机构地区：[1]山东政法学院山东省高校证据鉴识重点实验室,山东济南250014 [2]河南科技大学法医学院,河南洛阳471003

出　　处：《中国法医学杂志》2015年第3期231-234,共4页Chinese Journal of Forensic Medicine

基　　金：山东政法学院省高校证据鉴识重点实验室开放课题(KF201405);山东政法学院科研计划项目(2014F14B)

摘　　要：目的选择骨盆CT片上的特征指标,建立逐步回归方程,探讨其在法医学同一认定中的应用价值。方法收集160名不同被检查者骨盆CT影像片各1张,70名被检查者不同次骨盆CT影像片各2张。选择并测量骨盆CT片上的14项指标值,分别计算不同人随机分组相同测量指标的组间的差值,以及相同人不同次测量指标间的差值,运用二分类logistic逐步回归分析,建立各项指标的一元回归方程和多项指标的多元回归方程,并对方程进行盲测检验。结果建立的14个一元方程中同一认定的正确率在61.1%（骶骨耳状面后缘宽）~80.5%（第一骶椎平面左右髂骨前端间距）之间;建立的6个多元回归方程的正确率在80.5%~93.8%之间。盲测准确率为100%。结论本文在CT片上选择的14项特征指标可以用于同一认定,在使用时应尽可能选用多元指标以得到更准确的结果。Objective This research was aimed to find out the characteristic indexes of the pelvic CT for setting up regression equation, and to discuss its application in the consensus identification of forensic medicine. Methods Collect one slice of pelvic CT of 160 different inspectors and two slices of pelvic CT in different times of 70 inspectors. Select and measure the 14 indexes of the pelvic CT slices for calculating the differentiation of the same index in random grouping inspectors, and differentiation the same index of the different times in the same inspectors. Using the binary classification logistic regression analysis, set up a regression equation of one index and multiple regression equation of multiple indexes, and test the equation in blind. Results The accuracy of consensus identification in 14 regression equation was among 61.1% （ auricular surface of the sacrum trailing edge width） -80. 5% （ first sacral vertebra plane around ilium front spacing）; the accuracy of consensus identification in 14 regression equation was among 80. 5% -93. 8%. The accuracy of blind test was 100%. Conclusion 14 indexes of the pelvic CT slices could be used in consensus identification. In order to obtain more accurate results, multiple indicators should be chosen as far as possible.

关 键 词：法医人类学 骨盆CT片 同一认定 LOGISTIC回归分析 

分 类 号：D919[医药卫生—法医学]