广东东莞市男性青少年肩关节正位DR片骨龄研究

目的利用肩关节正位数字化X线片（DR片）上诸骨发育特点建立男性青少年骨龄评测数学模型。方法对广东东莞市310名15~22岁男性青少年拍摄肩关节正位DR片,根据锁骨肩峰端（X1）、肩胛骨肩峰端（X2）、喙突（X3）、肱骨近端（X4）及锁骨胸骨端（X5）骨骺发育分级评分,利用SPSS统计进行各骨骼发育评分与年龄间的相关分析、Fisher＇s线性两类判别分析及多元逐步回归分析。结果 X1~X5与生活年龄间的Pearson相关系数分别为0.496、0.571、0.559、0.702、0.807（P〈0.01）;16、17、18周岁判别分析综合判别率为87.2%~91.4%,年龄推断多元逐步回归方程准确率（±1岁）约79.8%。结论本研究建立的年龄判别分析方程及多元逐步回归方程,可为法医活体年龄鉴定提供了新的地区性标准和方法。     

利用中国汉族成人股骨上端CR片特征判定年龄

目的探讨中国汉族成人股骨上端CR片的特征随年龄变化的规律,以推断年龄。方法 277例中国汉族成人骨盆正位CR片,其中男性139例,年龄范围18-83岁,女性138例,年龄范围18-85岁,选取股骨上端的股骨头、股骨颈、大转子内的骨小梁及股骨上端髓腔高度四项指标,观察不同年龄各指标的变化进行统计分析。结果成人股骨头、股骨颈骨小梁变化和股骨上端髓腔高度是影响45岁以下年龄变化的重要因素,股骨大转子影响较小。四项指标对45岁以上年龄组均是影响年龄变化的重要因素。利用以上指标建立的回归方程能很好的反映成人股骨上端的增龄性变化,预测骨龄较为准确。结论利用成人股骨头、股骨颈、股骨大转子内骨小梁的改变及股骨上端髓腔高度变化可推断年龄。     

运用薄层CT扫描评估四川汉族青年锁骨胸骨端骨骼年龄

目的 运用薄层CT扫描探索四川汉族青年锁骨胸骨端骨骺发育状况及其与生活年龄的关系.方法 结合Schmeling等提出的骨发育分级法,并考虑本研究样本的年龄区间,将锁骨胸骨端骨骺发育分为4个等级.依据上述骨骺发育等级阅读565例15～25周岁青年胸部薄层CT片,并对锁骨胸骨端骨骺发育状况进行统计学描述性研究.结果 两性之...     

基于机器学习和锁骨胸骨端薄层CT的年龄推断研究

目的 采用Kellinghaus分级法对锁骨胸骨端薄层CT进行人工阅片分级,运用多种传统统计学方法以及机器学习方法构建青少年及成人早期年龄推断模型,探索机器学习技术在四川汉族人群年龄推断研究中的应用价值。方法 回顾性收集491例10～30岁个体的胸部薄层CT影像,参照Kellinghaus分级法对所收集样本进行阅片分级赋分。随机选取10%的数据作为测试集,其余数据作为训练集,综合构建多种青少年及成人早期年龄推断的传统统计学回归模型与机器学习模型,采用平均绝对误差值（Mean Absolute Error,MAE）对模型的性能进行评估。结果统计回归模型中效能最好的模型为三次回归模型,男性MAE值为1.34,女性MAE值为1.57;三种机器学习模型中,随机森林模型对男性的预测效能最好,MAE值为1.39;支持向量模型对女性的预测效能最好,MAE值为1.51。结论在锁骨胸骨端年龄推断模型的构建中,机器学习模型在年龄推断中的准确性有一定提升,但与传统统计学回归模型相比并无明显优势,机器学习方法在锁骨胸骨端年龄推断中的应用价值仍有待进一步探索研究。     

应用锁骨薄层CT容积重建影像推算四川汉族身高

目的探讨锁骨薄层CT三维容积重建影像(CT-VRT)在身高推算中的应用价值,以期更新我国西南地区身高推算数据,为鉴定实践提供技术支撑。方法应用300例锁骨薄层CT-VRT影像,测量并计算左侧锁骨长(LCL)、右侧锁骨长(RCL)、双侧锁骨均长(ACL),分析身高与LCL、RCL、ACL相关性及性别差异,建立身高推算线性回归方程,并对所建模型进行回代检验。结果男性、女性ACL和身高的相关系数分别为0.534、0.707;LCL和身高的相关系数分别为0.484、0.680;RCL和身高的相关系数分别为0.523、0.695。分别建立ACL、LCL、RCL男女性身高推算线性回归模型。回代检验示ACL方程男性身高推算平均绝对误差(MAD)为4.48cm,女性MAD为3.51cm;LCL方程男性身高推算MAD为4.60cm,女性MAD为3.64cm;RCL方程男性身高推算MAD为4.49cm,女性MAD为3.59cm。结论建立的推算身高的回归方程具有统计学意义,模型预测精度较高,表明利用薄层CT-VRT影像测量的锁骨可运用于四川汉族人群身高推算研究。但模型R2较小,因此在应用ACL进行身高推算时最好结合其他指标。     

挤压肩部致胸锁关节脱位1例

张某，男，20岁，学生。于1997年11月24日在协助推汽车出库的过程中被汽车挤压至车库墙壁而受伤，当时左肩直接接触于墙壁。入院查体：左胸锁关节处肿胀，左肩高于右肩约2．0cm，左锁骨胸骨端向前上方突起，向上移位约相当于一个锁骨胸骨端高度。胸部X线检查：左侧胸锁关节呈完全性前上脱位，锁骨胸骨端向上移位2．0cm。讨论胸锁关节是上肢与躯干之间的唯一关节，属微动关节，呈鞍状，由膨大的锁骨胸骨端、胸骨肋切迹、第一肋骨共同构成。关节囊坚韧，周围有胸锁前、后韧带分别加强于该关节的前、后壁及下壁。上述结构特点使胸锁关节比较稳…     

锁骨胸骨端CT扫描的法医学年龄推断

目的运用薄层胸部CT图像,观察锁骨胸骨端继发骨化中心随年龄增长的时序性变化,推测我国北方地区汉族青少年锁骨内侧骨骺发育程度在年龄增长过程中的分布特点。方法评估1369例11~27周岁的北方汉族青少年锁骨胸骨端二次骨化中心的发育情况,并将其进行分期。结果在所观察的数据中锁骨内侧骨化中心最早出现的年龄是12岁,开始融合的最小年龄是15岁,达到完全融合的最小年龄是19岁。在所有分期中,发展到Ⅱa和Ⅱb期时,两性间锁骨内侧骨化中心发育程度有差异,女性发育要早于男性。结论作为最晚愈合的骨化中心,锁骨内侧骨化中心对于法医学年龄的推断有重要的提示作用。     

X线摄影测量活体胫腓骨长度推算身高

目的应用计算机X线摄影（computer radiography,CR）测量活体胫腓骨长度,建立适合当代中国北方汉族人群胫腓骨长度推算身高的数学模型。方法在422例（男性207例．女性215例）活体山东汉族成年人的胫腓骨正位CR片上测量5项长度指标,同时测量其身高。通过各项测量指标与身高的线性回归分析、线性回归模型假设检验的方差分析以及偏回归系数的t检验．建立运用胫腓骨长度推算身高的回归方程;并对所建立的方程进行身高预测值与真实值的配对t检验、残差分析等方程的预测与诊断。结果共建立有统计学意义的一元回归方程35个一所建立的回归方程的相关系数在0．909～0．823范围内．回归方程的估计值标准误在2．48～3．87cm范围结论应用CR放射学方法测量胫腓骨长度推算身高．可以为法医学个人识别身高推算提供帮助,建立的回归方程适用于中国北方汉族人群胫腓骨的身高推算。     

挤压肩部致胸锁关节脱位1例

张某,男,20岁,学生.于1997年11月24日在协助推汽车出库的过程中被汽车挤压至车库墙壁而受伤,当时左肩直接接触于墙壁.入院查体:左胸锁关节处肿胀,左肩高于右肩约2.0 cm,左锁骨胸骨端向前上方突起,向上移位约相当于一个锁骨胸骨端高度.胸部X线检查:左侧胸锁关节完全性前上脱位,锁骨胸骨端向上移位2.0 cm.     

中国汉族女性青少年法医学活体骨龄推断数学模型的建立

目的建立推断中国汉族女性青少年活体骨龄的数学模型。方法摄取华中、华南及华东等地区的838名年龄介于11～20周岁正常女性青少年双侧锁骨胸骨端以及左侧肩、肘、腕、髋、膝、踝关节的X线片。依据青少年骨发育分级标准对24项骨骼发育指标进行阅片、分级,结合考虑身高、体质量及地区等影响因素．应用SAS8．1及SPSS11．0软件进行统计学处理,探索各指标与年龄的相关性。结果建立了我国汉族女性青少年利用锁骨胸骨端及6大关节骨骺闭合程度联合推断活体年龄的多元回归数学模型．推导出判定我国汉族女性青少年是否已满14、16和18周岁的Fisher’S两类判别分析方程。结论本研究所建立的判定活体年龄的数学模型丰富了活体年龄的法医学鉴定方法,有利于提高活体骨龄鉴定方法的科学性和结论的准确性。     

Determination of sex from the metacarpals in a Thai population

Determination of sex using metacarpals was carried out on a sample of 249 skeletons from a Thai population (154 males and 95 females), ranging in age from 19 to 93 years. Six measurements were taken on each metacarpal, namely maximum length, medio-lateral base width, antero-posterior base height, medio-lateral head width, antero-posterior head height and mid-shaft diameter. Binary logistic regression equations were calculated for determining sex from these measurements. All metacarpals from both sides produced at least one equation that correctly allocated the skeletons with 80% or greater accuracy. In a comparative test using only individuals with no missing measurements (n=196), the most accurate equations for each metacarpal on the right side had pooled allocation accuracies ranging from 85.2% to 89.3%, with the best equation based on three measurements from the 5th metacarpal. On the left side, the most accurate equations for each metacarpal ranged from 83.2% to 89.8% correct allocation, with the best equation based on three measurements of the 2nd metacarpal. When the allocation accuracy for each sex is considered in addition to the pooled accuracy, the best equations involve the same three measurements of the 5th metacarpal on the right side, but shift to three measurements of the first metacarpal on the left side, with a pooled accuracy of 88.3%. The results of this study suggest that metacarpals can be used quite reliably for sexing in forensic contexts in Thailand.     

Age estimation from quantitation of features of "chest plate" X-rays

This study presents a method for estimating the age at death from the quantitation of roentgenologic features of X-ray films of chest plates obtained during routine autopsies. Multiple linear regression analysis allows estimation of coefficients of regression of features on known age-at-death individuals. The regression equation can be used in turn for age estimation of an unknown age-at-death individual. The accuracy of age estimation is about +/- 8.4 years (standard error) which is in the range of previously published macroscopic methods, though the present method is much faster and simpler.     

Accuracy of BR Regression Equation for Impacted Teeth in Age Estimation in Haryana Population of India

Abstract:  Dental cementum is a vital tissue that demonstrates continuous apposition throughout the life of a tooth. Positive correlation has been reported between the coronal displacement of cementum and age in impacted teeth and BR regression equation has been proposed. The present study is an attempt to calculate accuracy of BR regression equation in Haryana population of India. The study was conducted on 20 impacted mandibular third molars extracted from healthy patients, aged between 22 and 32 years. The bucco-lingual ground sections were prepared and the distance between the edges of enamel and cementum were measured with micrometer attached to a light microscope and BR regression equation was applied for age estimation. Although the estimated age with BR regression equation was less than chronological age it was not statistically significant. Hence, the BR regression equation can be used in medicolegal as well as in forensic odontology for age estimation.     

A test of sex determination from measurements of chest radiographs

Sexually dimorphic distinctions within the human thoracic area may include morphological as well as metric differences in the sternum and 4th rib. This research assesses the validity of a set of previously published measurements from chest radiographs and their use in contemporary forensic situations. The chest plates from 130 adult individuals of a known sample undergoing medico-legal post-mortem examination were examined at autopsy. Thoracic radiographs were taken using a Faxitron cabinet X-ray machine at 40 kV using Kodak Diagnostic Film Ready Pack X-Omat TL. Measurements were taken to the nearest millimetre using a sliding calliper. Logistic regression analysis of measurements of the sternum and 4th rib was undertaken to determine sex. Using 4th rib width and sternal area, sex was predicted at an accuracy of 95.8% for males and 90.3% for females.     

放射影像学方法测量活体女性颈椎推算身高

目的测量颈椎高度建立四川汉族女性人群的身高推算方程，为法医学个人识别提供帮助。方法应用计算机X线摄影放射学方法，在209例四川汉族女性颈椎CR片上测量C3～C7椎体高度，并准确测量身高。按年龄是否大于45岁进行分组，对各颈椎的测量数据与身高进行线性回归分析，建立颈椎推算身高的回归方程。结果共建立一元回归方程36个，多元回归方程12个，所有回归方程经线性回归模型假设检验，均有统计学意义（P〈0．05）。所建立的回归方程的复相关系数（r）在0．628～0．280范围内，回归方程的标准误在4．03～5．03cm范围内。结论本研究所建立的回归方程，可用于四川汉族女性颈椎的身高推算。     

Chest deflection tolerance to blunt anterior loading is sensitive to age but not load distribution

Ninety-three human cadaver tests are used in the development of thoracic injury risk functions with consideration of age and restraint condition. Linear logistic regression models are developed with the set of potential predictors including the maximum chest deflection, the age of the cadaver at death, gender, and the loading condition on the anterior thorax: blunt hub (41 tests), seat belt (26 tests), air bag (12 tests), and combined belt-and-bag (14 tests). Predicted outcomes were the probability of any rib fractures (onset of injury) and the probability of greater than six rib fractures (severe injury). The analysis shows that the injury risk function was not dependent on the loading condition, but was strongly dependent on age. A significant injury risk model with good ability to discriminate injury from non-injury tests (P < 0.0001, chi-square = 21.49, area under receiver operator characteristic curve (ROC) = 0.867, Kruskal's Gamma = 0.732) is presented using only maximum chest deflection and cadaver age as predictors of injury risk. The 50% risk of any rib fractures is found to occur at 35% chest deflection for a 30-year-old, but at 13% deflection for a 70-year-old. The 50% risk of severe injury is shown to occur at 33% chest deflection for a 70-year-old, but at 43% for a 30-year-old.     

Validation of a radiographic method to establish new fetal growth standards: radio-anatomical correlation

In forensic medicine, specialists might face difficulties when estimating age at death from fetal remains. Depending on the state of preservation, this age assessment is essentially based on the diaphyseal size of long bones. In a previous work, for the measurement of fetal femoral ossified shafts, we already established a simple and reliable method using a radiographic protocol. Since we previously stated that radiographic measurement values were closer to real anatomical size than ultrasonographic ones, in the present study we decided to check the importance of the difference between radiographic and anatomical measurements. Therefore, we dissected 30 pairs of fetal femurs and compared the difference between the two kinds of measurements (in percentages). This difference seemed to be slight (4.027%), but it was large enough to entail significant differences (p < 0.001). In order to provide a correction factor for radiographic measurements, we established a linear regression formula, which was tested on another sample of 30 pairs of dissected femurs. As a consequence of the good results, we improved the linear regression using a powerful statistical tool: the bootstrap. Finally, we obtained a simple equation that allowed us to figure out the real anatomical size with an R2 of 99% and a mean relative difference of 0.153% (with a standard error of 0.252 mm, and therefore a 95% confidence interval with limits of -0.35 and 0.657 mm). This difference did not entail any significant differences (p = 0.498), and therefore, we concluded that with the proposed correction, radiographic measurements can easily be used by forensic specialists in their daily tasks or to establish new growth standards in order to best fit their population of interest.     

Age estimation using carpals: study of a Slovenian sample to test Cameriere's method

Carpals are often used as age indicators. In a recent study, Cameriere et al. studied the use of the ratio between the total area of carpal bones and epiphyses of the ulna and radius (Bo) and carpals (Ca) as age indicators. The present study, of a sample of 158 Slovenian children and adolescents aged between 6 and 16 years, focused on analysing the best regression for age estimation. The regression model yielded the following equation: age=-3.411+0.942 g+20.927(Bo/Ca), and explained 91.6% of total variance (R(2)=0.916). The median of the absolute values of residuals (observed age minus predicted age) was 0.09 years, with a quartile deviation of 0.786 years, and a standard error of estimate of 0.658 years. Comparisons between the previous equation referring to Slovenian children and the equivalent linear equation proposed by Cameriere et al. did not reveal any significant differences between the intercepts and slopes of the two linear models. These results suggested a common regression model for both Italian and Slovenian samples. The common regression model, describing age as a linear function of gender and Bo/Ca ratio, yielded the following linear regression formula: age=-2.907+0.408 g+20.757(Bo/Ca). This model explained 86% of total variance (R(2)=0.86). The median of the absolute values of residuals (observed age minus predicted age) was 0.02 years, with a quartile deviation of 1.02 years and a standard error of estimate of 0.96 years.