Age estimation of living South African individuals : a multifactorial model

Abstract

Age estimation in living individuals around the legal age of 18 years remains a difficult challenge, with limited options available. In this study third molar development was used, along with the novel method of anterior inferior vertebral ring apophysis development, to assess the age of living individuals and the probability of being 18 years. For third molar development, panoramic radiographs of 705 white and 563 black South Africans were scored using a 10 stage scoring system. Vertebral apophysis development of C2, C3, and C4 of 496 white and 478 black South Africans were assessed from cephalometric radiographs, using a four-stage scoring system. Likelihood values were determined for individuals in each sex and population group being 18 years, based on developmental stages. For apophysis development, the median ages for attainment of stages 0, 1, and 2 were below the 18-year threshold for all ancestry and sex groups, while stage 3 was also below this threshold in some groups. For third molar development, black South African individuals consistently matured earlier than white South African individuals, while for most of the stages the opposite was true for apophysis development. Differences between the sexes were also noted for third molar, but not for vertebral apophysis development. These age indicators were also combined by using a generalised linear model (GLM). The combined sample comprised of 165 females and 122 males aged between 15 and 18 years. Four additional models were obtained from data sets only containing data for third molar and cervical ring apophysis development respectively. The performance of all the models were quantified and compared using the Akaike information criterion (AIC) as an estimator of the relative quality of the statistical models and the prediction error as a mean square error value. The best performance resulted from third molar development, although the vertebral data adds a component related to skeletal development which may better reflect biological reality. These results show that cervical vertebral apophysis development is a valuable, novel addition to the assessment of age in living individuals. Both these methods are easy to use and can be assessed from standard and routinely used radiographic images. The developed models need to be sex and ancestry specific, as clear differences were noted.