Age estimation based on dental characteristics observed on panoramic radiographs in the Japanese population: A...

In addition to conventional forensic age-estimation methods, panoramic radiographs offer an opportunity to develop new approaches based on observable dental characteristics. This study aimed to develop regression models for age estimation in the Japanese population and compare their accuracy with that of models developed for the Korean population. Panoramic radiographs from 420 Japanese individuals (20-90 years old, 214 male and 206 female participants) were used in this study. Each tooth was classified into nine codes (V, virgin; X, missing; I, impacted; D, defect; R, residual root; T, root canal treatment; F, restoration; P, prosthesis; and L, dental implant). Five dental codes (X, T, F, P, and L) selected for the regression analysis were evaluated for their correlations with chronological age, and four regression models were developed for each of the nine combination patterns to estimate age in the Japanese population through a comparative evaluation of the model performance. The accuracy of the estimated ages was compared with that obtained by applying the regression models derived from the Korean population to the Japanese datasets. We found that the accuracy of the age-estimation regression model for the maxillary and mandibular premolars and molars in the overall combination model achieved an adjusted root mean square error of 12.77 years, comparable to that of the whole dentition model (13.23 years). Japanese datasets yielded slightly different age estimates when Korean-derived regression models were applied. These findings suggest the potential applicability of dental treatment codes in panoramic radiographs for forensic age estimation.