Age estimation using Hounsfield unit values from computed tomography of proximal femur trabecular: A validation study in the Thai population

Abstract

Background: Age is one of the biological profiles that are important in human identification, not only of unknown remains but also forensic clinical patients who need their identification confirmed, such as marginal patients who needs to obtain a national identification (ID) card.
Objective: The present study aimed to validate the age estimation equation of Ford JM, et al. in the Thai population to test the accuracy of each method when applied to a Thai sample.
Methods: The study uses computed tomography (CT) scans and the resultant 3D models of 48 male and 48 female patients aged 18-to 88-years-old to measure their trabecular bone density of the proximal femur in Hounsfield units (HU) and compares the actual age and age calculated by the age estimation equation.
Results: The actual age and age calculated using the combined sex, male and female equations are significantly different (P < 0.05). The bias and inaccuracy range are 2.8 - 37.4, the combined sex age estimation equation gives an overall accuracy of 16.7% ± 5 years, with the highest accuracy of 83.3% found in females aged 60 - 69 years.
Conclusion: The equation of Ford JM, et al. cannot be in age prediction among the Thai population and this suggests that inter-population variation does exist. Further study with a larger sample, using cadavers, might give better results for age estimation among the Thai population.