Multivariate cumulative probit for age estimation using ordinal categorical data

Radiographic age estimation based on degenerative changes of vertebrae

Age estimation is an important component of decedent identification. When assessing adult remains, anthropologists frequently use gross examination of skeletal elements, such as clavicles, ribs, and pubic symphyses. For fleshed bodies, this requires the removal of these elements and maceration prior to analysis. A new method was developed using radiographic imaging to estimate age from degenerative changes of the lower thoracic and upper lumbar vertebrae. This technique will complement anthropological age estimation methods in young and middle-aged adults and may serve as a stand-alone method for older individuals. Digital radiographs from 240 medical examiner cases were evaluated. The sample included 120 females and 120 males between the ages of 18 and 101 years. A 3-phased scoring system was used for the target vertebrae. Transition analysis was conducted on binned average scores and a Bayesian approach was used to assign age intervals. At the 90% credible interval, individuals in Bin 1 were under 36 years of age while those in Bin 3 were over 47 years of age. Individuals in Bin 2 showed too much age variation to be informative. No significant differences were found between males and females. These findings will be especially useful in the age estimation of older adults and may eliminate the need for skeletal sampling in medicolegal cases where advanced degenerative changes are radiographically observed in the lower thoracic and/or upper lumbar vertebrae. This method was developed for use on fleshed individuals but may also be applicable to skeletonized remains.

Age estimation by modified Suchey-Brooks method using three-dimensional reconstructed CT images of Chinese Han population

The method proposed by Suchey-Brooks for adult age estimation based on the surface morphology of the pubic symphysis has been widely accepted. The applicability of the method varies considerably in different populations. The present study established a virtual reference sample and aimed to develop population-specific criteria that can be used for age estimation in different skeletal samples. First, The dry bone specimens from 100 individuals were compared with their corresponding three-dimensional (3D) reconstruction model and showed high inter-method agreement (k = 0.743-0.811), suggesting that the virtual bone model and physical bone specimens have comparable performances in describing the surface morphology of the pubic symphysis. We retrospectively collected clinical computed tomography (CT) data from 895 Chinese patients to create a virtual reference sample of the pubic symphysis. Based on the original Suchey-Brooks method, each of the 895 reference samples was assigned a phase, for each sex and phase, data on the mean age, standard deviation, and 95% age range of the corresponding sample were obtained, which was then used as the "method modified for Chinese" (modified method) and compared to the "SB method". Compared to the SB method, modified method had a lower inaccuracy in dry bones for males over 35 years and females over 45 years, in dry bone CT test sample for males over 55 years and females over 45 years, and in postmortem CT test sample for males over 35 years and females over 55 years. The modified method can improve the accuracy of age estimation for older samples over 40 years. It has shown considerable reliability when applied as a population-specific criterion, but its accuracy is still not sufficient, and caution is needed when using it.

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.

Adult Skeletal Age-at-Death Estimation through Deep Random Neural Networks: A New Method and Its Computational Analysis

Age-at-death assessment is a crucial step in the identification process of skeletal human remains. Nonetheless, in adult individuals this task is particularly difficult to achieve with reasonable accuracy due to high variability in the senescence processes. To improve the accuracy of age-at-estimation, in this work we propose a new method based on a multifactorial macroscopic analysis and deep random neural network models. A sample of 500 identified skeletons was used to establish a reference dataset (age-at-death: 19–101 years old, 250 males and 250 females). A total of 64 skeletal traits are covered in the proposed macroscopic technique. Age-at-death estimation is tackled from a function approximation perspective and a regression approach is used to infer both point and prediction interval estimates. Based on cross-validation and computational experiments, our results demonstrate that age estimation from skeletal remains can be accurately (~6 years mean absolute error) inferred across the entire adult age span and informative estimates and prediction intervals can be obtained for the elderly population. A novel software tool, DRNNAGE, was made available to the community.

Interpreting error in the estimation of skeletal growth profiles from past populations: An example demonstrating skeletal growth in historic African American communities

OBJECTIVES

The study of growth in the past is a critical component of bioarcheological analyses. However, our understanding of growth in the past is subject to a number of methodological challenges. This study aims to model the skeletal growth of past populations by considering the challenges associated with the data collection process and the challenges associated with the age estimation procedures.

MATERIALS AND METHODS

We use skeletal remains from two historic African American cemeteries in the American South to model femoral diaphyseal length-for-age. We estimate the age of each individual using dental development techniques and present growth curves as both a product of the maximum likelihood (MLE) age estimate and the estimated posterior age distribution. Growth was compared against a reference sample from the University of Colorado Child Research Council Study.

RESULTS

The results of our analyses showed that femoral diaphyseal length in two historic African American communities is small-for-estimated age as compared to a modern reference sample. However, the magnitude and characterization of this difference is variable when taking into account the broader posterior age distribution.

DISCUSSION

Both samples may be small-for-age due to physiological stress associated with racism, inequality, and the compounding effects of early urbanization. However, the interpretation of growth in the past is muddled when considering the relationship between the study sample and the reference sample, when accounting for uncertainty in the age estimation procedure, and the error-inducing steps taken during the data collection process. Future interpretation of skeletal growth in the past must include a full account of the possible sources of error in order to present an accurate representation of growth.

Conditional independence assumption and appropriate number of stages in dental developmental age estimation

When estimating the age of an individual it is critical that 1) age ranges are as narrow as possible while still capturing the true age of the individual with an acceptable frequency, and 2) this frequency is known. When multiple traits are used to produce a single age estimate, the simplest practice is to assume that the traits are conditionally independent from one another given age. Unfortunately, if the traits are correlated once the effect of age is accounted for, the resulting age intervals will be too narrow. The frequency at which the age interval captures the true age of the individual will be decreased below the expected value to some unknown degree. It is therefore critical that age estimation methods that include multiple traits incorporate the possible correlations between them. Moorrees et al. (1963) [1] scores of the permanent mandibular dentition from 2607 individuals between 2 and 23 years were used to produce and cross-validate a cumulative probit model for age estimation with an optimal number of stages for each tooth. Two correction methods for covariance of development between teeth were tested: the variance-covariance matrix for a multivariate normal, and the Boldsen et al. (2002) [2] ad-hoc method. Both correction methods successfully decreased age interval error rates from 21% to 23% in the uncorrected model to the expected value of 5%. These results demonstrate both the efficacy of these correction methods and the need to move away from assuming conditional independence in multi-trait age estimation.

A probability model for assessing age relative to the 18-year old threshold based on magnetic resonance imaging of the knee combined with radiography of third molars in the lower jaw

OBJECTIVE

This study aims to generate a statistical model based on magnetic resonance imaging of the knee and radiography of third molars in the lower jaw, for assessing age relative to the 18-year old threshold.

METHODS

In total, 58 studies correlating knee or tooth development to age were assessed, 5 studies for knee and 7 studies for tooth were included in the statistical model. The relation between the development of the anatomical site, based on a binary system, and age were estimated using logistic regression. Separate meta-populations for knee and tooth were generated from the individual based data for men and women. A weighted estimate of probabilities was made by combining the probability densities for knee and tooth. Margin of errors for males and females in different age groups and knee and tooth maturity were calculated within the larger framework of transition analysis using a logit model as a base. Evidentiary values for combinations of knee and tooth maturity were evaluated with likelihood ratios.

RESULTS

For males, the sensitivity for the method was calculated to 0.78 (probability of correctly classifying adults), the specificity 0.90 (probability of correctly classifying minors), the negative predictive value 0.80 (proportion identified minors are minors) and the positive predictive value 0.89 (proportion identified adults are adults) indicating a model better at identifying minors than adults. The point at which half the female population has reached closed knee lies before the 18-year threshold, adding the knee as an indicator lowers specificity and increases sensitivity. The sensitivity when using tooth as an indicator for females is 0.24 and specificity 0.97, signifying few minors misclassified as adults but also a low probability of identifying adults. The negative predictive value for women when using tooth as the sole indicator is 0.56 and positive predictive value 0.88. Probabilities were calculated for males and females assuming a uniform age distribution between 15 and 21years. The calculated margin of error of minors classified as adults in a population between 15 and 21 years with the model was 11% for males and 12% for females. Further, the evidentiary value as well as margin of error vary for different combinations of knee and tooth maturity.

CONCLUSION

The statistical model based on the combination of MRI knee and radiography of mandibular third molars is a valid method to assess age relative to the 18-year old threshold when applied on males and of limited value in females.

Magnetic resonance imaging for forensic age estimation in living children and young adults: a systematic review

The use of MRI in forensic age estimation has been explored extensively during the last decade. The authors of this paper synthesized the available MRI data for forensic age estimation in living children and young adults to provide a comprehensive overview that can guide age estimation practice and future research. To do so, the authors searched MEDLINE, Embase and Web of Science, along with cited and citing articles and study registers. Two authors independently selected articles, conducted data extraction, and assessed risk of bias. They considered study populations including living subjects up to 30 years old. Fifty-five studies were included in qualitative analysis and 33 in quantitative analysis. Most studies had biases including use of relatively small European (Caucasian) populations, varying MR approaches and varying staging techniques. Therefore, it was not appropriate to pool the age distribution data. The authors found that reproducibility of staging was remarkably lower in clavicles than in any other anatomical structure. Age estimation performance was in line with the gold standard, radiography, with mean absolute errors ranging from 0.85 years to 2.0 years. The proportion of correctly classified minors ranged from 65% to 91%. Multifactorial age estimation performed better than that based on a single anatomical site. The authors found that more multifactorial age estimation studies are necessary, together with studies testing whether the MRI data can safely be pooled. The current review results can guide future studies, help medical professionals to decide on the preferred approach for specific cases, and help judicial professionals to interpret the evidential value of age estimation results.

Dental and Skeletal Imaging in Forensic Age Estimation: Disparities in Current Approaches and the Continuing Search for Optimization

Medical imaging for forensic age estimation in living adolescents and young adults continues to be controversial and a subject of discussion. Because age estimation based on medical imaging is well studied, it is the current gold standard. However, large disparities exist between the centers conducting age estimation, both between and within countries. This review provides an overview of the most common approaches applied in Europe, with case examples illustrating the differences in imaging modalities, in staging of development, and in statistical processing of the age data. Additionally, the review looks toward the future because several European research groups have intensified studies on age estimation, exploring four strategies for optimization: (1) increasing sample sizes of the reference populations, (2) combining single-site information into multifactorial information, (3) avoiding ionizing radiation, and (4) conducting a fully automated analysis.

Age estimation in forensic anthropology: methodological considerations about the validation studies of prediction models

There is currently no clear consensus on how to calculate, express, and interpret the error when validating methods for age estimation in forensic anthropology. For this reason, it is likely that researchers are commonly drawing erroneous or confusing conclusions about the existence of population differences or the need to design new and increasingly complex estimation methods. In recent years, many researchers have highlighted these limitations. They propose new lines of research focused on the use of rigorous statistics and new technologies for the development of methods for estimating age. Our main objective in this study is to contribute to the strengthening of these novel ideas, for which we show the existing empirical evidence about the inadequacy of some age estimation methods in calculating, expressing, and interpreting the errors obtained. With this aim, a total of 500 simulations have been performed, in which hypothetical research teams develop and validate methods for age estimation. The data employed in this study was obtained from the "Centers for Disease Control and Prevention (CDC) Growth Charts: United States" released in 2000. The charts relate age with height, weight, and head circumference of US male children. Five learning algorithms have been employed as age estimators. We have performed three experiments in which the following aspects have been analyzed: frequency with which "negative" results can be obtained in the validation studies; which are the most appropriate criteria to compare and select the age estimation methods; and what analysis should be employed to carry out the validation studies. The results show possible errors in the interpretation of validation studies as a consequence of the confusion of statistical concepts. To conclude, we made a proposal of "good practices" for the correct calculation, expression, and interpretation of the error when validating age estimation methods in forensic anthropology.

A Study on the Asymmetry of the Human Left and Right Pubic Symphyseal Surfaces Using High-Definition Data Capture and Computational Shape Methods

The pubic symphysis is among the most commonly used bilateral age indicators. Because of potential differences between right and left sides, it is necessary to investigate within-individual asymmetry, which can inflate age estimation error. This study uses 3D laser scans of paired pubic symphyses for 88 documented White males. Scan data are analyzed by numerical shape algorithms, proposed as an alternative to traditional visual assessment techniques. Results are used to quantify the within-individual asymmetry, evaluating if one side produces a better age-estimate. Relationships between the asymmetry and advanced age, weight, and stature are examined. This analysis indicates that the computational, shape-based techniques are robust to asymmetry (>80% of paired differences are within 10 years and >90% are within 15 years). For notably more asymmetric cases, differences in estimates are not associated with life history factors. Based on this study, either side can be used for age-at-death estimation by the computational methods.

Reliability of age estimation from iliac auricular surface in a subactual Chilean sample

The aim of this study was to test the reliability of Osborne et al. (2004) [8] age estimation based on the iliac auricular surface of the ilium. We selected 172 skeletons with documented sex and age-at-death and established six uniformly distributed age intervals for analysis. ANOVA was employed to assess the association of the documented age with sex and the auricular surface classification. We employed Bonferroni post-hoc tests to find any statistical differences across documented ages within each phase of Osborne's criteria. While the ANOVA showed a significant association between the documented age and the auricular surface classification, post-hoc tests found that phases 2 and 3 were the only consecutive phases with significant differences. We argue that a lack of statistical significance between consecutive phases undermines the reliability of this method for forensic purposes especially in middle adults. It may be possible that the collapse of intermediate phases would increase the accuracy of this method.

An overview of age estimation in forensic anthropology: perspectives and practical considerations

CONTEXT

Information on methods of age estimation in physical anthropology, in particular with regard to age-at-death from human skeletal remains, is widely available in the literature. However, the practicalities and real challenges faced in forensic casework are not always highlighted.

OBJECTIVES

To provide a practitioner's perspective, regarding age estimation in forensic anthropology (both in the living as well as the dead), with an emphasis on the types of cases, the value of such work and its challenges and limitations.

METHODS

The paper reviews the current literature on age estimation with a focus on forensic anthropology, but it also brings the author's personal perspective derived from a number of forensic cases.

RESULTS AND CONCLUSIONS

Although much is known about what methods to use, but not always how to apply them, little attention has been given in the literature to the real practicalities faced by forensic anthropologists, for example: the challenges in different types of scenarios; how to report age estimations; responsibilities; and ethical concerns. This paper gathers some of these aspects into one overview which includes the value of such work and the practical challenges, not necessarily with the methods themselves, but also with regard to how these are applied in the different cases where age estimation is required.