Age estimation in children by measurement of open apices in teeth with Bayesian calibration approach

• Age estimation by radiological analysis has wide applications in scientific fields.
• Bayesian calibration method appears to be suitable for assessing age.
• Forensic age estimation is essential in resolving a variety of legal questions.
• Age assessment is useful in diagnosis and in orthodontic treatment.

Age estimation from teeth by radiological analysis, in both children and adolescents, has wide applications in several scientific and forensic fields. In 2006, Cameriere et al. proposed a regression method to estimate chronological age in children, according to measurements of open apices of permanent teeth.Although several regression models are used to analyze the relationship between age and dental development, one serious limitation is the unavoidable bias in age estimation when regression models are used.The aim of this paper is to develop a full Bayesian calibration method for age estimation in children according to the sum of open apices, S, of the seven left permanent mandibular teeth.This cross-sectional study included 2630 orthopantomographs (OPGs) from healthy living Italian subjects, aged between 4 and 17 years and with no obvious developmental abnormalities. All radiographs were in digital format and were processed by the ImageJ computer-aided drawing program. The distance between the inner side of the open apex was measured for each tooth. Dental maturity was then evaluated according to the sum of normalized open apices (S).Intra- and inter-observer agreement was satisfactory, according to an intra-class correlation coefficient of S on 50 randomly selected OPGs. Mean absolute errors were 0.72 years (standard deviation 0.60) and 0.73 years (standard deviation 0.61) in boys and girls, respectively. The mean interquartile range (MIQR) of the calibrating distribution was 1.37 years (standard deviation 0.46) and 1.51 years (standard deviation 0.52) in boys and girls, respectively. Estimate bias was βERR = −0.005 and 0.003 for boys and girls, corresponding to a bias of a few days for all individuals in the sample. Neither of the βERR values was significantly different from 0 (p > 0.682).In conclusion, the Bayesian calibration method overcomes problems of bias in age estimation when regression models are used, and appears to be suitable for assessing both age and age distribution in children according to tooth maturity.