Evaluation of Standards for Sex Estimation in Black South Africans Using Measurements Derived from a Three-Dimensional Computed Tomography (3DCT) Scan

Abstract

There is a growing need to identify humans whose identity at death, time and circumstances of death are questionable. This is due to the rising human activities that have the potential to distort nature and lead to natural disasters, criminal activities as well other circumstances that require the need for human identification. Different demographic information such as sex, stature, age, and ancestry is often kept by anthropologists for identification purposes. However, in unusual circumstances (for example, natural disasters, suicide or accidents) with massive casualties, dismembering of the human body, and dead and decaying bodies, the need to utilise a more intrinsic means of identification becomes an utmost necessity. A correct sex estimation can help unravel other anthropological parameters because when the sex of an individual is known it narrows the identification possibilities of a subject by 50% due to the reduction of potential matching identities. South Africa has a multi- dimensionally high crime rate, dispute over identity of a deceased, cases homicides and so on, and its largest distinct population group (Black South Africans) is most affected by such occurrences. Several authors have in the past used measurements of the skull and postcranial elements for sex discrimination among South Africans. Such attempts have largely been made using direct osteometry by using the Vernier calliper and osteometric board to obtain relevant measurements from post-mortem human skeletal remains that are usually housed in osteological collections across the country. This study, therefore, attempted to overcome the drawbacks of cadaveric-based bone measurements by using a three-dimensional computed tomography (3-DCT) to obtain population-specific data from a contemporary Black South African population group. Measurements from CT records of 350 crania (M=175; F=175), 320 pelves (M=160; F=160), 320 femora (M=160; F=160) and 97 humeri (M=64; F=33) of Black South Africans whose age ranged between 18 to 60 years were utilised. The CT scans were housed in the Radiology Department of the Charlotte Maxeke Johannesburg Academic Hospital as Digital Imaging and Communications in Medicine (DICOM) files. The Xiris and IntelliSpace software was used to reconstruct the images into two-dimensional (2-D) and three-dimensional (3-D) forms from which measurements were taken. The reliability of previously formulated equations for sex estimation from measurements of the crania, humeri, pelves, and femora of the same population group were evaluated using data collected from an independent sample in the current study. The average accuracies in correct sex assignment ranged between 50.0% and 72.0% for the crania, it was vii 73.5% for the humeri, ranged from 50.0% to 85.0% for the pelves and 50.0% to 75.0% for the femora. This necessitated the formulation of new discriminant function equations for sex estimation from measurements of the cranium, humerus, pelvis and femur. The average accuracies for correct sex estimation from this study ranged from 77.4% to 82.90% for the crania, 83.5% to 90.7% for the humeri, 86.4% to 91% for the pelves and 79.3% to 84.3% for the femora. Cross validation of these equations using data from an independent sample showed a range of average accuracies of 78.0% to 90.0% for the cranium, 90.0% to 100% for the pelvis and 82.5% to 100% for the femur. These indicate that the functions formulated from the CT- derived measurements in the current study are more reliable in the estimation of sex of contemporary Black South Africans compared to those from cadaveric-derived functions.