Dental age estimation using radiographs: Towards the best method for Sri Lankan children

Abstract Research supports the need for regionally and ethnically specific validated data as the reference base for age estimation techniques. This retrospective study evaluated the accuracy of three dental age estimation methods; Demirjian et al. (1973), Willems et al. (2001), and Blenkin and Evans (2010), for use in Sri Lanka for medico-legal purposes. Panoramic radiographs of 688 Sri Lankan children ranging in age from 8.00 to 16.99 years were used to determine their appropriateness to a Sri Lankan population. The mean age and standard deviations (±SD) were calculated separately for males and females of each age group. Paired t -test and mean absolute errors (MAE) were calculated to compare the calculated dental age (DA) with the chronological age (CA) across the nine age groups. The results revealed the mean CA of the entire sample was 12.38 ± 2.68 years, while the mean DA calculated using the Blenkin and Evans method was 11.83 ± 2.20 years, using the Demirjian et al. method was 12.57 ± 2.53 years, and using the Willems et al. method was 11.99 ± 2.43 years. The Demirjian et al. method consistently overestimated the age of males except in the 2 groups aged over 15 years, whereas the Blenkin and Evans method consistently underestimated the age except for the 11.00–12.99 age range. The method of Willems et al. produced DA quite close to CA up until 12.99 years of age, and then underestimated the age for all higher age groups. In females, the Demirjian et al. method consistently overestimated the age up until 13.99 years and then underestimated the higher age groups, while the Willems et al. method underestimated the age in all age groups except 10.00–10.99 years. The Blenkin and Evans method also consistently underestimated the age except in the 10.00–10.99 and 12.00–12.99-year age groups. The percentages of either overestimation or underestimation calculated for ±0.5 years of the true age were 41.0% for the Blenkin and Evans method, 42.8% for the Demirjian et al. method and 49.1% for the Willems et al. method. In conclusion, while all three methods could be applicable in the estimation of dental age for medico-legal purposes, the Willems et al. method appears to be more appropriate in overall measures for the Sri Lankan reference sample, up to the age of 12.99 years.