四川汉族青少年肘关节数字X线影像骨龄特征研究

目的观测四川地区青少年肘关节数字X影像的骨发育特征规律,建立肘关节推断年龄的回归方程。方法拍摄608例8～18岁青少年进行肘关节数字X线片,按性别及年龄分组,选取肘关节6个骨化中心作为观测指标,对发育特征进行分级、赋值。采用SPSS 15.0统计软件对各观测值进行描述性统计学分析,建立推算年龄的回归方程。结果得出男女性青少年肘关节继发骨化中心出现、闭合的年龄范围、平均年龄以及95%置信区间等数据。一元回归方程的复相关系数(R)范围:男性0.874～0.914,女性0.883～0.954;年龄估计值标准误(SE)范围:男性0.934～1.101岁,女性0.561～0.883岁;多元回归方程R值及SE值:男性为0.946、0.750岁,女性为0.964、0.500岁。所有方程均经统计学检验与诊断,具有较好的适用性。结论本文选择的肘关节指标发育特征与年龄有相关性,其中肱骨内上髁、桡骨小头及肱骨小头骨骺均呈高度显著相关,建立的回归方程可用于相关人群年龄推断,且多元回归方程准确性更高。

Skeletal age estimation in Sichuan region Han teenagers with elbow radiographs

Objective To establish the mathematical model of skeletal age estimation by analyzing the X-ray imaging regularities of the elbow epiphyses in Sichuan region Han normal teenagers.Methods The radiographs of the elbow joints aged from 8~18 in Sichuan region Han Population were taken to investigate X-ray signs of epiphysis growth of the elbow joints.6 mark points were selected and X-ray staging criteria were enacted.Then correlation and regression analysis between various index and the age were carried on to establish the regression equations of skeletal age estimation using SPSS 15.0 statistical software.Results The appearance and the closure of the elbow epiphyses were collected.In the group of the single regression equations,the correlation coefficient(R) of ranging was 0.914~0.874 for males and 0.954~0.883 for females.The standard errors values(SE) of the years of the predicted ranging was 0.934~1.101 years for males,which was 0.561 ~ 0.883 year for female.As for the combining multiple regression group,R was 0.946 for males and 0.964 for females.SE was 0.750 year for males and 0.500 year for females.All of the equations were strictly statistical tested and diagnosed,and with good applicability.Conclusion The mark points of the elbow epiphyses we selected,especially the medial epicondyle,the radial head and the humeral head epiphyses,are highly correlated to ages.The established regression mathematical models may have potential to evaluate the skeletal age,the multiple regression equation may be better than regression equation established with a single index.