Human age estimation combining third molar and skeletal development
Human age estimation combining third molar and skeletal development
复制标题
DOI：
10.1007/s00414-011-0639-5
复制
发表时间：
2012-03-01
影响因子：
2.1
通讯作者：
Willems, G.
中科院分区：
文献类型：
--
作者：
Thevissen, P. W.;Kaur, J.;Willems, G.
根据第三磨牙发育的年龄估计方法获得的预测区间很宽，可以通过将这些牙齿观察与年龄相关的骨骼信息相结合来减少。因此，在头影测量X线片上，搜索最准确的年龄估计骨骼变量和相关配准方法，并将其添加到回归模型中，年龄作为响应，第三磨牙阶段作为解释变量。在496张（283 M; 213 F）头影测量X线片数据集的试点设置中，验证了Baccetti等人（2005）（BA）、Seedat等人（2005）（SE）、Caldas等人（2007）和Rai等人（2008）（RA）的技术。在主要研究中，收集了460例（208例女性，224例男性）年龄在3 - 26岁之间的个体的数据，在同一天拍摄了正位正位片和头颅片。在曲面断层片上，使用Gleiser和Hunt（1955）描述并经科勒（1994）（GH）修改的评分系统记录左侧第三磨牙的发育情况。在头颅X线片上，根据BA和SE技术记录颈椎发育。回归模型，以年龄为响应和GH分数作为解释变量，拟合的数据。然后，将BA、SE和BA + SE的信息分别添加到该模型中。从所有获得的模型中，计算决定系数和均方根误差。纳入基于BA和SE技术的头影X线照片信息，可以将使用GH技术从全景X光片中获得的年龄解释方差提高48%。纳入头影测量BA + SE信息略微改善了先前的结果（+1%）。通过在GH模型中加入BA、SE和BA + SE信息，RMSE分别下降了1.93、1.85和2.03年。SE技术允许临床上最快和最简单的颈椎发育程度配准。因此，除了第三磨牙发育之外，对颈椎发育进行分类的技术选择最好是SE技术。
The wide prediction intervals obtained with age estimation methods based on third molar development could be reduced by combining these dental observations with age-related skeletal information. Therefore, on cephalometric radiographs, the most accurate age-estimating skeletal variable and related registration method were searched and added to a regression model, with age as response and third molar stages as explanatory variable. In a pilot set up on a dataset of 496 (283 M; 213 F) cephalometric radiographs, the techniques of Baccetti et al. (2005) (BA), Seedat et al. (2005) (SE), Caldas et al. (2007) and Rai et al. (2008) (RA) were verified. In the main study, data from 460 (208 F, 224 M) individuals in an age range between 3 and 26 years, for which at the same day an orthopantogram and a cephalogram were taken, were collected. On the orthopantomograms, the left third molar development was registered using the scoring system described by Gleiser and Hunt (1955) and modified by Kohler (1994) (GH). On the cephalograms, cervical vertebrae development was registered according to the BA and SE techniques. A regression model, with age as response and the GH scores as explanatory variable, was fitted to the data. Next, information of BA, SE and BA + SE was, respectively, added to this model. From all obtained models, the determination coefficients and the root mean squared errors were calculated. Inclusion of information from cephalograms based on the BA, as well as the SE, technique improved the amount of explained variance in age acquired from panoramic radiographs using the GH technique with 48%. Inclusion of cephalometric BA + SE information marginally improved the previous result (+1%). The RMSE decreased with 1.93, 1.85 and 2.03 years by adding, respectively, BA, SE and BA + SE information to the GH model. The SE technique allows clinically the fastest and easiest registration of the degree of development of the cervical vertebrae. Therefore, the choice of technique to classify cervical vertebrae development in addition to third molar development is preferably the SE technique.