Machine Learning Supported the Modified Gustafson's Criteria for Dental Age Estimation in Southwest China

Abstract

Adult age estimation is one of the most challenging problems in forensic science and physical anthropology. In this study, we aimed to develop and evaluate machine learning (ML) methods based on the modified Gustafson's criteria for dental age estimation. In this retrospective study, a total of 851 orthopantomograms were collected from patients aged 15 to 40 years old. The secondary dentin formation (SE), periodontal recession (PE), and attrition (AT) of four mandibular premolars were analyzed according to the modified Gustafson's criteria. Ten ML models were generated and compared for age estimation. The partial least squares regressor outperformed other models in males with a mean absolute error (MAE) of 4.151 years. The support vector regressor (MAE = 3.806 years) showed good performance in females. The accuracy of ML models is better than the single-tooth model provided in the previous studies (MAE = 4.747 years in males and MAE = 4.957 years in females). The Shapley additive explanations method was used to reveal the importance of the 12 features in ML models and found that AT and PE are the most influential in age estimation. The findings suggest that the modified Gustafson method can be effectively employed for adult age estimation in the southwest Chinese population. Furthermore, this study highlights the potential of machine learning models to assist experts in achieving accurate and interpretable age estimation.

Keywords: Age determination by teeth; Forensic dentistry; Gustafson’s criteria; Machine learning; Orthopantomograms.

Fig. 1

Heatmap of the error of all 15 models in the test set. a Males. b Females. The vertical axis is each sample and age, and the horizontal axis is the model. Green means underestimation of age and red means overestimation of age

Fig. 2

Bland and Altman graphs of the difference in years between chronological age and dental age. a The partial least squares regression of males; b the support vector regression of females. Negative values indicate that dental age is lower than chronological age

Fig. 3

The interpretation of the optimal models in males and females. a The importance ranking of features in descending order according to the mean (|SHAP value|) in males; b the importance ranking of features with stability and interpretation in males; c the importance ranking of features in descending order according to the mean (|SHAP value|) in females; d the importance ranking of features with stability and interpretation using the optimal model in females. The higher the SHAP value of a feature is given, the higher age of the patient. The red part in the feature value represents a higher feature value