西南地区汉族男性CT颅面径线与身高的关系

目的通过CT测量西南地区汉族男性颅面径线,建立该地区汉族男性颅面径线与身高的回归模型,为法医人类学研究积累数据。方法收集273例西南地区汉族男性头部CT资料,确定颅面部7条径线,运用图像后处理软件进行多平面重建和容积再现,对所选径线进行测量。对各测量指标与身高的相关性运用SPSS 21.0进行分析,建立推断身高的回归方程,并重新选取50例样本代入该数学模型进行检验。结果建立了7条径线的线性回归方程(P0.05);一元线性回归方程相关系数为0.190~0.439,标准估计误差(SEE)为4.597~5.023 cm;多元线性回归方程复相关系数为0.494~0.524,SEE为4.418~4.458 cm。回代检验得y=83.959+3.589 x_6+2.573 x_2,±1SEE准确率最高,为30%;y=72.646+3.316 x6+1.586 x_2+1.553 x_4+2.211 x_3,±2SEE准确率最高,为92%。结论本研究所选取的7条径线与身高均有明显的直线相关性,所建立的多元回归方程适用于推断西南地区汉族男性身高。

Relationship between Body Height and Craniofacial Lines Measured by CT in Southwest Han Males

Objective To establish regression m odel betw een craniofacial lines and body height by m ea-suring craniofacial lines in Southw est H an m ales using C Tand to accum ulate data for the study of foren-sic anthropology. Methods H ead C Tdata of 273 H an m ales in Southw est w ere collected and 7 cranio-facial lines w ere determ ined. M ultiplanar reconstruction and volum e rendering w ere perform ed by im age post-processing softw are and the selected lines w ere m easured. The relationship betw een each m easuring indicator and body height w as analyzed using SPSS 21.0 softw are. The regression equation of body height estim ation w as established and 50 sam ples w ere selected again and put into the m athem atics m odels to verify its accuracy. Results The linear regression equations of 7 lines w ere established (P<0.05). The correlation coefficients of the unary linear regression equations w ere 0.190-0.439 and the standard errors of the estim ate (SEE) w ere 4.597-5.023 cm . The correlation coefficients of the m ultiple linear regression equation w ere 0.494-0.524 and the SEEw ere 4.418-4.458 cm . The return tests show ed that the highest ±1SEEaccuracy of the m ultiple regression equation:y=83.959+3.589 x6+2.573 x2, w ere 30%;and the highest ±2SEEaccuracy of the m ultiple regression equation: y=72.646+3.316 x6+1.586 x2+1.553 x4+2.211 x3, w ere 92% . Conclusion There is significant linear correlation betw een 7 selected lines and the stature in this study, and the plural linear regression equation established could be applied for estim ating the stature of Southw est H an m ales.