Linear densitometry and digital image processing of proximal femur radiographs: implications for archaeological and...

Age- and sex-related patterns of proximal femur trabecular bone loss have previously been used to establish radiographic reference standards for estimating age at death for human skeletal remains. Such standards are of interest to both anthropologists and forensic scientists. However, osteopenia as a physiological phenomenon is dependent on numerous genetic, environmental, and cultural factors. Thus, while general age- and sex-related trends can be clearly observed for trabecular bone loss, such patterns also demonstrate marked variation among individuals of both sexes at all ages. Moreover, clinical evidence shows that rates of bone loss are not steady but episodic, and that radiographically "normal" (i.e., young adult) patterns of trabecular bone architecture can also exist in femora of older individuals, particularly within samples of African origin. In this study, adult proximal femur radiographs were used to explore patterns of age- and sex-related proximal femur cancellous bone involution among a sample of 66 African-American individuals from the Terry collection (33 males and 33 females), ranging in age from 19 to 71 years. The proximal femur radiographs of these subjects were analyzed by digital image processing (DIP), and the results were compared to those obtained by laser linear densitometric analyses (LDA) previously performed on the same series (Macchiarelli et al., 1987). Results of LDA and DIP analyses indicate (a) more pronounced bone density decrease in females; (b) sex- and site-specific structural patterns of proximal femur trabecular bone loss; (c) a high level of individual variability, in which predicted age deviated from real age by as much as 22.26 (males) and 30.78 years (females); (d) a moderate linear correlation with age for all the variables analyzed; and (e) an average discrepancy between known age and predicted age (measured by root mean squared residual values) of 10.34 (males) and 12.71 years (females) for the most satisfactory DIP analysis parameter (recorded at the center of the femoral neck in the equalized images). With regard to estimating adult age at death from radiographic images of proximal femora, it is argued that the amount of useful age information reported for this criterion has been overestimated.