Bayesian networks of age estimation and classification based on dental evidence: A study on the third molar mineralization

Machine learning assessment of dental age classification based on cone-beam CT images: a different approach

OBJECTIVES

Machine learning (ML) algorithms are a portion of artificial intelligence that may be used to create more accurate algorithmic procedures for estimating an individual's dental age or defining an age classification. This study aims to use ML algorithms to evaluate the efficacy of pulp/tooth area ratio (PTR) in cone-beam CT (CBCT) images to predict dental age classification in adults.

METHODS

CBCT images of 236 Turkish individuals (121 males and 115 females) from 18 to 70 years of age were included. PTRs were calculated for six teeth in each individual, and a total of 1416 PTRs encompassed the study dataset. Support vector machine, classification and regression tree, and random forest (RF) models for dental age classification were employed. The accuracy of these techniques was compared. To facilitate this evaluation process, the available data were partitioned into training and test datasets, maintaining a proportion of 70% for training and 30% for testing across the spectrum of ML algorithms employed. The correct classification performances of the trained models were evaluated.

RESULTS

The models' performances were found to be low. The models' highest accuracy and confidence intervals were found to belong to the RF algorithm.

CONCLUSIONS

According to our results, models were found to be low in performance but were considered as a different approach. We suggest examining the different parameters derived from different measuring techniques in the data obtained from CBCT images in order to develop ML algorithms for age classification in forensic situations.

A collection of idioms for modeling activity level evaluations in forensic science

This paper presents a collection of idioms that is useful for modeling activity level evaluations in forensic science using Bayesian networks. The idioms are categorized into five groups: cause-consequence idioms, narrative idioms, synthesis idioms, hypothesis-conditioning idioms, and evidence-conditioning idioms. Each category represents a specific modeling objective. Furthermore, we support the use of an idiom-based approach and emphasize the relevance of our collection by combining several of the presented idioms to create a more comprehensive template model. This model can be used in cases involving transfer evidence and disputes over the actor and/or activity. Additionally, we cite literature that employs idioms in template models or case-specific models, providing the reader with examples of their use in forensic casework.

Solving the inverse problem of post-mortem interval estimation using Bayesian Belief Networks

Bayesian Belief Networks (BBNs) can be applied to solve inverse problems such as the post-mortem interval (PMI) by a simple and logical graphical representation of conditional dependencies between multiple taphonomic variables and the observable decomposition effect. This study is the first cross-comparison retrospective study of human decomposition across three different geographical regions. To assess the effect of the most influential taphonomic variables on the decomposition rate (as measured by the Total Decomposition Score (TDS)), decomposition data was examined from the Forensic Anthropology Research Facility at the University of Tennessee (n = 312), the Allegheny County Office of the Medical Examiner in Pittsburgh, US (n = 250), and the Crime Scene Investigation department at Southwest Forensics in the UK (n = 81). Two different BBNs for PMI estimations were created from the US and the UK training data. Sensitivity analysis was performed to identify the most influential parameters of TDS variance, with weaker variables (e.g., age, sex, clothing) being excluded during model refinement. The accuracy of the BBNs was then compared by additional validation cases: US (n = 28) and UK (n = 10). Both models conferred predictive power of the PMI and accounted for the unique combination of taphonomic variables affecting decomposition. Both models had a mean posterior probability of 86% (US) and 81% (UK) in favor of the experimental hypothesis (that the PMI was on, or less than, the prior last known alive date). Neither the US nor the UK datasets represented any cases below 'moderate' support for the value of PMI evidence. By applying coherent probabilistic reasoning to PMI estimations, one logical solution is provided to model the complexities of human decomposition that can quantify the combined effect of several uncertainties surrounding the PMI estimation. This approach communicates the PMI with an associated degree of confidence and provides predictive power on unknown PMI cases.

Facial biotype classification for orthodontic treatment planning using an alternative learning algorithm for tree augmented Naive Bayes

BACKGROUND

When designing a treatment in orthodontics, especially for children and teenagers, it is crucial to be aware of the changes that occur throughout facial growth because the rate and direction of growth can greatly affect the necessity of using different treatment mechanics. This paper presents a Bayesian network approach for facial biotype classification to classify patients' biotypes into Dolichofacial (long and narrow face), Brachyfacial (short and wide face), and an intermediate kind called Mesofacial, we develop a novel learning technique for tree augmented Naive Bayes (TAN) for this purpose.

RESULTS

The proposed method, on average, outperformed all the other models based on accuracy, precision, recall, [Formula: see text], and kappa, for the particular dataset analyzed. Moreover, the proposed method presented the lowest dispersion, making this model more stable and robust against different runs.

CONCLUSIONS

The proposed method obtained high accuracy values compared to other competitive classifiers. When analyzing a resulting Bayesian network, many of the interactions shown in the network had an orthodontic interpretation. For orthodontists, the Bayesian network classifier can be a helpful decision-making tool.

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.

Comparison of different machine learning approaches to predict dental age using Demirjian's staging approach

CONTEXT

Dental age, one of the indicators of biological age, is inferred by radiological methods. Two of the most commonly used methods are using Demirjian's radiographic stages of permanent teeth excluding the third molar (Demirjian's and Willems' method). The major drawbacks of these methods are that they are based on population-specific conversion tables and may tend to over- or underestimate dental age in other populations. Machine learning (ML) methods make it possible to create complex data schemas more simply while keeping the same annotation system. The objectives of this study are to compare (1) the capacity of ten machine learning algorithms to predict dental age in children using the seven left permanent mandibular teeth compared to reference methods and (2) the capacity of ten machine learning algorithms to predict dental age from childhood to young adulthood using the seven left permanent mandibular teeth and the four third molars.

METHODS

Using a large radiological database of 3605 orthopantomograms (1734 females and 1871 males) of healthy French patients aged between 2 and 24 years, seven left permanent mandibular teeth and the 4 third molars were assessed using Demirjian's stages. Dental age estimation was then performed using Demirjian's reference method and various ML regression methods. Two analyses were performed: with the 7 left mandibular teeth without third molars for the under 16 age group and with the third molars for the entire study population. The different methods were compared using mean error, mean absolute error, root mean square error as metrics, and the Bland-Altman graph.

RESULTS

All ML methods had a mean absolute error (MAE) under 0.811 years. With Demirjian's and Willems' methods, the MAE was 1.107 and 0.927 years, respectively. Except for the Bayesian ridge regression that gives poorer accuracy, there was no statistical difference between all ML tested.

CONCLUSION

Compared to the two reference methods, all the ML methods based on the maturation stages defined by Demirjian were more accurate in estimating dental age. These results support the use of ML algorithms instead of using standard population tables.

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 assessment by Demirjian's development stages of the third molar: a systematic review

OBJECTIVES

Radiographic evaluation of the wisdom teeth (third molar) formation is a widely used age assessment method for adolescents and young adults. This systematic review examines evidence on the agreement between Demirjian's development stages of the third molar and chronological age.

METHODS

We searched four databases up until May 2016 for studies reporting Demirjian's stages of third molar and confirmed chronological age of healthy individuals aged 10-25 years. Heterogeneity test of the included studies was performed.

RESULTS

We included 21 studies from all continents except Australia, all published after 2005. The mean chronological age for Demirjian's stages varied considerably between studies. The results from most studies were affected by age mimicry bias. Only a few of the studies based their results on an unbiased age structure, which we argue as important to provide an adequate description of the method's ability to estimate age.

CONCLUSION

Observed study variation in the timing of Demirjian's development stages for third molars has often been interpreted as differences between populations and ethnicities. However, we consider age mimicry to be a dominant bias in these studies. Hence, the scientific evidence is insufficient to conclude whether such differences exist.

KEY POINTS

• There is significant heterogeneity between studies evaluating age assessment by Demirjian's third molar development. • Most of the studies were subject to the selection bias age mimicry which can be a source of heterogeneity. • Presence of age mimicry bias makes it impossible to compare and combine results. These biased studies should not be applied as reference studies for age assessment.