Human age estimation from lower-canine pulp volume ratio based on Bayes’ theorem with modern Japanese population as prior distribution

Abstract

Pulp volume decreases throughout life owing to secondary dentin deposition. Here, we present a Bayesian approach for human age estimation on the basis of the pulp volume ratio from lower-canine teeth. We measured the pulp and tooth volumes of 363 subjects (209 males, 154 females) of known age based on three-dimensional imageries from microcomputed tomography scans. Pulp volume ratio was defined as pulp/tooth volume within the root portion (PVRrt) and its reduction with age was modeled by simple statistical assumptions to produce a likelihood function of PVRrt by fitting the model to the observed data. Following Bayes’ theorem, we obtained the probability density function (PDF) of estimated age for a given PVRrt, for males, females, and gender unknowns; we used the modern Japanese population as of 2012 as prior age distribution. The PDF of estimated age provided the mean and 50%, 70%, and 90% prediction intervals, based on which we compared the present age- estimation method with that of Brooks and Suchey (Human Evolution, 5: 227–238), which is based on pubic symphysis metamorphosis. We could not find any advantages of the PVRrt method compared to the Suchey–Brooks method for the case of males or gender unknowns as far as the prediction intervals were concerned. However, for females, the PVRrt method offered comparable precision to the Suchey–Brooks method. Taking into account other advantages such as sustainability of the material and continuous property of the age-indicator value, we conclude that the PVRrt method is useful in forensic cases, especially those involving female victims.