Auricular Surface Aging: Comparing Two Methods that Assess Morphological Change in the Ilium with Bayesian Analyses

Age estimation from iliac auricular surface using Bayesian inference and principal component analysis: a CT-based study in an Indian population

Age estimation constitutes one of the pillars of human identification. The auricular surface of the ilium presents as a durable and robust structure within the human skeletal framework, capable of enabling accurate age estimation in older adults. Amongst different documented auricular age estimation methods, the Buckberry-Chamberlain method offers greater objectivity through its component-based approach. The present study aimed to test the applicability of the Buckberry-Chamberlain method in an Indian population through a CT-based examination of the auricular surface. CT scans of 435 participants undergoing CT examinations following the advice of their treating physicians were scrutinized for different age-related auricular changes. Three of the five morphological features described by Buckberry-Chamberlain could be appreciated on CT scans, and thus further statistical analysis was restricted to these features. Transition analysis coupled with Bayesian inference was undertaken individually for each feature to enable age estimation from individual features, while circumventing age mimicry. A Bayesian analysis of individual features yielded highest accuracy percentages (98.64%) and error rates (12.99 years) with macroporosity. Transverse organization and apical changes yielded accuracy percentages of 91.67% and 94.84%, respectively, with inaccuracy computations of 10.18 years and 11.74 years, respectively. Summary age models, i.e. multivariate age estimation models, derived by taking this differential accuracy and inaccuracy into consideration yielded a reduced inaccuracy value of 8.52 years. While Bayesian analysis undertaken within the present study enables age estimation from individual morphological features, summary age models appropriately weigh all appreciable features to yield more accurate and reliable estimates of age.

Mixed cumulative probit: a multivariate generalization of transition analysis that accommodates variation in the shape, spread and structure of data

Biological data are frequently nonlinear, heteroscedastic and conditionally dependent, and often researchers deal with missing data. To account for characteristics common in biological data in one algorithm, we developed the mixed cumulative probit (MCP), a novel latent trait model that is a formal generalization of the cumulative probit model usually used in transition analysis. Specifically, the MCP accommodates heteroscedasticity, mixtures of ordinal and continuous variables, missing values, conditional dependence and alternative specifications of the mean response and noise response. Cross-validation selects the best model parameters (mean response and the noise response for simple models, as well as conditional dependence for multivariate models), and the Kullback-Leibler divergence evaluates information gain during posterior inference to quantify mis-specified models (conditionally dependent versus conditionally independent). Two continuous and four ordinal skeletal and dental variables collected from 1296 individuals (aged birth to 22 years) from the Subadult Virtual Anthropology Database are used to introduce and demonstrate the algorithm. In addition to describing the features of the MCP, we provide material to help fit novel datasets using the MCP. The flexible, general formulation with model selection provides a process to robustly identify the modelling assumptions that are best suited for the data at hand.

Age-at-death estimation from the auricular surface of the ilium: A test of a sex-specific component method

Accurate age-at-death estimation is important for both paleodemographic studies and forensic casework. Although the auricular surface of the ilium is a well-validated skeletal indicator for aging studies, problems persist with identifying features that estimate age accurately in older individuals. This study tests the utility of one method, developed by Igarashi et al. (2005), which claims to estimate age more accurately in older individuals using a presence/absence scoring system for 13 auricular surface traits. Four hundred (400) individuals, aged 16-93 years, from the Hamann-Todd Collection were examined to test the performance of Igarashi et al.'s method in a North American sample. Pearson's product-moment correlation tests were performed for both the overall method and individual traits to assess correlation with chronological age. Eleven of the 13 traits showed statically significant correlations with chronological age, and nine were found to have higher correlations than originally reported. The method showed a tendency toward negative bias (i.e., a tendency to under-age individuals, particularly in the older age range). Models for both males and females and full and reduced models developed by Igarashi et al. were tested; the sex-pooled full model performed best, and the female full model performed most poorly. Although this method did not have significantly higher accuracy rates in a North American sample than other auricular surface methods, unique traits identified by Igarashi et al. did correlate with chronological age. In future studies, these traits should be investigated using different scoring systems (e.g., character states), as they show utility for aging research.

Performance of three mathematical models for estimating age-at-death from multiple indicators of the adult skeleton

The mathematical method which will achieve the most accurate and precise age-at-death estimate from the adult skeleton is often debated. Some research promotes Bayesian analysis, which is widely considered better suited to the data construct of adult age-at-death distributions. Other research indicates that methods with less mathematical complexity produce equally accurate and precise age-at-death estimates. One of the advantages of Bayesian analysis is the ability to systematically combine multiple indicators, which is reported to improve the age-at-death estimate. Few comparisons exist between Bayesian analysis and less complex mathematical models when considering multiple skeletal indicators. This study aims to evaluate the performance of a Bayesian approach compared to a phase-based averaging method and linear regression analysis using multiple skeletal indicators. The three combination methods were constructed from age-at-death data collected from 330 adult skeletons contained in the Raymond A Dart and Pretoria Bone Collections in South Africa. These methods were tested and compared using a hold-out sample of 30 skeletons. As is frequently reported in literature, a balance between accuracy and precision was difficult to obtain from the three selected methods. However, the averaging and regression analysis methods outperformed the Bayesian approach in both accuracy and precision. Nevertheless, each method may be suited to its own unique situation-averaging to inform first impressions, multiple linear regression to achieve statistically defensible accuracies and precisions and Bayesian analysis to allow for cases where category adjustments or missing indicators are necessary.

The accuracy of age estimation using transition analysis in the Hamann-Todd collection

OBJECTIVES

Reconstructing demography of past populations using skeletal data is challenging when analyzing adults because the process of biological aging does not always reflect the individual's chronological age. A proposed solution to address the limitations of traditional age estimation methods is transition analysis (TA), a multifactorial method of age estimation. However, despite its methodological refinement, TA has varying degrees of accuracy when applied to different known-age skeletal samples. This study assesses TA's accuracy by comparing age estimates to known age at death in the Hamann-Todd Collection.

MATERIALS AND METHODS

We contrasted the maximum likelihood age estimates generated by the ADBOU program to the known ages of 221 individuals. The absolute error was calculated for the entire sample, and compared between sex and ancestry.

RESULTS

The mean absolute error in the sample is 11.6 (SD = 10.3) years, with white individuals' errors (14.1 years) being significantly higher than black individuals' (9.1 years; p < 0.001). No significant differences were found between sexes (p = 0.621). A weak to moderate positive correlation was found between known age and absolute error for white males (R = 0.607; p < 0.001), white females (R = 0.509; p < 0.001), and black males (R = 0.371; p = 0.006). The accuracy of TA age estimates varied when each anatomical region was analyzed independently, but the combination of all three anatomical regions yielded the most accurate age estimates.

DISCUSSION

These findings further support that TA accuracy depends on the prior distribution used and, in the Hamann-Todd Collection, the accuracy for white individuals is more influenced by this limitation than when black individuals are analyzed.

Sacroiliac joint morphologic changes from infancy to adulthood

BACKGROUND CONTEXT

Report of sacroiliac morphology changes during growth is limited in the literature and the interest of such morphology and its consequence for surgery is increasing.

PURPOSE

Aims of this work are (1) to anatomically define the sacroiliac joint (SIJ), and (2) to assess the influence of growth on the sacroiliac morphology and the pelvic parameters.

STUDY DESIGN

Forty-nine young subjects from 6 months to 18 years old (y/o) and 20 adults aged from 18 to 50 y/o were selected from our institutional patient database.

METHODS

They underwent a computed tomography (CT) examination on a 128-MDCT (GE Healthcare Optima CT660). Transverse CT image datasets were reconstructed, anonymized, and segmented with ITK-SNAP. Landmarks and surfaces were selected and a SIJ orientation analysis was performed using costumed Python scripts.

RESULTS

The subjects were divided into four groups: infants (1.9±1 y/o), children (6.9±1.7 y/o), adolescents (13.7±1.8 y/o), and adults (27.3±5.6 y/o). Differences between SIJ orientation were found significant between young subject groups for synovial sacrum SIJ orientation (p<.001) and iliac total SIJ orientation (p=.036). Both orientations of younger subjects were found significantly different from the adult group (p<.035). SIJ synovial sacrum and iliac total orientations correlated significantly with age (p<.03). All orientations correlated with pelvic incidence (p<.04) except for synovial sacrum SIJ orientation (p=.2). No gender or symmetrical differences were found significant in any group.

CONCLUSIONS

Morphologic definition of the SIJ confirmed the independency of the gender during growth. Such results will be beneficial for the analysis and management of vertebral pathology.

The Use of Informative Priors in Bayesian Modeling Age-at-death; a Quick Look at Chronological and Biological Age Changes in the Sacroiliac Joint in American Males

The aim of this study is to examine how well different informative priors model age-at-death in Bayesian statistics, which will shed light on how the skeleton ages, particularly at the sacroiliac joint. Data from four samples were compared for their performance as informative priors for auricular surface age-at-death estimation: (1) American population from US Census data; (2) county data from the US Census data; (3) a local cemetery; and (4) a skeletal collection. The skeletal collection and cemetery are located within the county that was sampled. A Gompertz model was applied to compare survivorship across the four samples. Transition analysis parameters, coupled with the generated Gompertz parameters, were input into Bayes' theorem to generate highest posterior density ranges from posterior density functions. Transition analysis describes the age at which an individual transitions from one age phase to another. The result is age ranges that should describe the chronological age of 90% of the individuals who fall in a particular phase. Cumulative binomial tests indicate the method performed lower than 90% at capturing chronological age as assigned to a biological phase, despite wide age ranges at older ages. The samples performed similarly overall, despite small differences in survivorship. Collectively, these results show that as we age, the senescence pattern becomes more variable. More local samples performed better at describing the aging process than more general samples, which implies practitioners need to consider sample selection when using the literature to diagnose and work with patients with sacroiliac joint pain.

Sex and age at death estimation from the sternal end of the fourth rib. Does Íşcan's method really work?

The fourth rib has been used commonly in order to estimate age at death and even sex in skeletal remains but many often, Iscan's estimates do not adjust to the real age of the individual. Population specific references for sex and age-at-death estimation from the sternal end of the fourth rib are presented for a contemporary Mexican sample. A total of 504 ribs with known sex and age from a morgue sample were studied (444 males, 60 females, 17 to 92 years old). The height and breadth of the sternal end of the rib were sexually dimorphic (p = .000), and allowed a correct sex assignment in 73.3% to 84% of the cases from univariate and multivariate discriminant functions. With regard to age-at-death estimation, the morphological changes summarized by the phases of the sternal end of the fourth rib are correlated with known age in this sample (Spearman's Rho, p = .000). However, the original age intervals tend to underestimate age at death and inaccuracy increases with phase scored in males. Descriptive statistics for rib phase are provided for males and females, and new age-at-death estimates based on transition analysis and Bayesian statistics are provided for the male sample. The test of universally applied methods and the development of population specific references is an important task for forensic anthropology around the world.

The auricular surface as age indicator in a modern Greek sample: A test of two qualitative methods

The auricular surface is often found very well preserved, thus age-related changes in this anatomical area can be important for any set of human remains that require identification under different taphonomic conditions. This study tests the Buckberry and Chamberlain (2002) and Schmitt et al. (2005) methods in predicting the age of individuals in a documented sample from Crete, Greece. Both methods were used to record changes on the auricular surface in a mixed-sex sample of 74 individuals, directly as well as through photographs, by two independent observers. Cohen's kappa and intra class correlation coefficients (ICC) were used in order to assess inter-observer and intra-observer agreement. Results showed than none of the methods predicted age with sufficient accuracy, as high error rates were recorded. The Schmitt et al. (2005) method performed better, mainly because the age ranges it uses are broader. Scoring through photographs does not seem to introduce bias in predicting age, as demonstrated by the high intra-observer agreement rates. Inter-observer agreement was also high. The low intra- and inter-observer error rates suggest that the poor performance of both methods in the Cretan sample is not due to a lack of clarity in the description of the morphological changes recorded on the auricular surface; rather it should be attributed to a poor correlation between these changes and age at death in our material.