The scientific analysis of the age of majority using computed tomography

Abstract

This thesis examines three developing anatomical sites, two skeletal and one dental, in an attempt to devise a method for more accurately estimating the age of unknown age individuals in the age range 15-25 years. Individuals in this age cohort have been the subject of relatively little research since there are few anatomical sites still undergoing development beyond the age of 15 years, and all of these have been reported to be quite variable. The three age markers chosen for this study were the spheno-occipital synchondrosis, the medial clavicular epiphysis and the third molar tooth, being the three most commonly examined development sites for this age cohort. All of these were examined using Multi Detector Computed Tomography (MDCT) scanning as the imaging modality, and the sample was derived from the large database of post-mortem full body CT scans of deceased individuals located at the Victorian Institute of Forensic Medicine, Melbourne, Australia. The study sample consisted of 1006 full body CT scans of individuals aged between 15 and 25 years. Each age marker was examined and a developmental score was applied according to accepted scoring methodology developed by others. Each anatomical site was individually assessed for its relationship with chronological age, and it was found that the third molar tooth and medial clavicle followed expected trends as to variability and developmental timing. The spheno-occipital synchondrosis was found to fuse earlier than expected in this population and had essentially completed development by the age of 17 years in both sexes. An investigation was undertaken to determine if combining the third molar and medial clavicle into one multiple regression analysis, the “multi-factorial approach” would result in an improvement in precision and accuracy of age estimates when compared to age estimations calculated using each age marker individually. It was found that the multi-factorial approach reduced age ranges by approximately half, thus providing higher precision and with no loss of accuracy at the 95% CI level. Further investigations were conducted to assess the effect of socio-economic status on development in the Australian population, with findings suggesting that there is no appreciable effect on development between the highest and lowest socio-economic groups in this country. Ancestry was also investigated, with the conclusion that the high level of racial admixture in this population renders the attribution of any particular racial group to an individual prone to error and somewhat irrelevant. Left/right fluctuating asymmetry was also examined, with results suggesting that in approximately 3.4% of individuals the difference in development between left and right sides will have an appreciable effect upon age estimation calculations, in some cases as much as 3 years. Recommendations to deal with this asymmetry are discussed. The multifactorial model presented in this thesis provides a means to more accurately estimate the age of individuals within the Australian population in this age cohort. This method is applicable to living people as well as the deceased. Future research needs to be directed at assessing the applicability of these formulae to members of other relevant populations, particularly those populations from which Australia receives a high number of asylum seekers, many of whom are in this age cohort and their status as adult or child is unknown.