Dental development has shown to be one of the more uniform processes in humans and it strongly correlates with the actual (chronological) age of an individual. Dental radiographs are often used to estimate dental age as it can be used in living individuals, is economical, easy to access and reliable [14]. This study investigated the reliability of the London Atlas in age estimation of the Australian population. This was carried out by measuring the difference in chronological and estimated age of the individual, which indicated either an overestimation or underestimation of chronological age.

Using the London Atlas, there was an underestimation in the age for males and overestimation in the age for females. This supports studies using Demirjian’s methods, where the overestimation and underestimation of ages were explained by slight discrepancies in the dental development between males and females [11, 15]. However, there was no significant difference in the accuracy of the London Atlas in estimating ages of males or females, and this coincides with results obtained from the study of the London Atlas in the Hispanic and Thai population [16, 17].

The London Atlas was generally found to overestimate the ages of children in the Australian population. This was an average overestimation of 0.36 years, which equates to approximately four and a half months. This difference of four and a half months of age would not result in significant undesirable consequences, when applied to real world circumstances. The overestimation bias was statistically significant in several of the younger age groups, indicating that the atlas may be more inaccurate and overestimates more so in the younger population of Australian children. This slight overestimation was consistent with the results obtained when the London Atlas was applied to the Thai population [17]. Furthermore, in this study, children aged 10 and 11 had the largest over estimation (1.11 years and 1.05 years respectively). Recent studies on the Hispanic and Brazilian children population gave similar results. When the London Atlas was applied, Hispanic children aged 11 to 12 and Brazilian children aged 11 had one of the largest overestimations [16, 18].

The average overestimation of 0.36 years from this study is consistent with the study conducted in Hispanic children using The London Atlas, where an average overestimation of 0.35 years was obtained [16]. This result also concurs with the findings of the London Atlas in the Portuguese population, although the overestimated amount was one month, which was lesser than that in this study [19]. Additionally, The London Atlas during this study proved similar estimation to Demirjian’s methods previously tested in various Australian populations [11, 12]. Flood et al. used Demirjian’s methods including more recent modified methods and found average age overestimation ranging from 0.31 to 0.61 years in a South Australian population [12]. This method is usually found to be more time consuming due to more in depth analysis of each individual tooth.

This study utilised an almost equal distribution of males (n = 97) and females (n = 96). Additionally, a significant number of radiographs were initially obtained to ensure a higher probability of members from the population to be included in the dataset. These radiographs were also filtered according to a strict inclusion and exclusion criteria, which produced 193 suitable radiographs.

This study had quite a few limitations. While the radiographs had strict inclusion and exclusion criteria, the quality of the radiographs could have varied. These images were taken on different machines in the facility and taken by different personnel. This non-standardised quality could have affected the interpretation of the developmental stages. Besides, all three intra-examiner reliability scores were consistently slightly lower than that of the inter-examiner scores which can possibly be due to factors such as examiner fatigue or a learning effect that may have influenced the intra-examiner results. While this is a potential limitation, the values were still 0.800 and above and hence it does not detract from the overall reliability and validity of the assessment process used for this study.

The London atlas age estimation method could also be limited by the variation in the development of third molars. However, to this date, it is debatable whether the inclusion of third molars in age estimation affects the accuracy of the results [20]. It is evident that third molars vary in development, morphology and positioning, and could potentially introduce more error in predicting one’s age [20]. Several studies even proposed that the rate of third molars developing, differs across different countries and ethnic groups [20, 21]. However, these studies concluded that the small differences in third molar development had little impact on age estimations due to the large standard deviation of developmental stages in each age group [20, 21]. Additionally, this population-specific study targeting solely the Australian population, reduced the error of possible cross-population variability. The study of the London Atlas for age estimation in Portugese population also concluded that third molars should be used in age estimation as the results yielded are as reliable as using a whole set of teeth [19]. Another limitation of the London atlas is that it estimates the age to the precision of the midpoint of each year (e.g. 6.5 years), whereas other studies including Demirijian’s methods, produces point estimates of age. However, the London Atlas has proven to be superior in accuracy when compared with most point estimate methods [5, 11].

The difference between the dental age and chronological age might also not be attributed to the shortcomings of the respective dental age estimation methods. Similar to other biological processes of the human species, dental development also has a certain amount of variation [14, 16]. Therefore, this difference might not be a limitation of the respective age estimation method but a limitation to using tooth eruption patterns to estimate chronological age [16].

It is likely that technologies like machine learning and artificial intelligence (AI) will play a larger role in forensic age estimation moving forward, as other studies also suggest [22]. However, we believe it is crucial to first grasp the foundational principles of this area of research before relying on advanced technologies, so as to better interpret the outcomes of automated methods. The London Atlas is emerging amongst many other age estimation methods. It is easily accessible in various languages and has comprehensive illustrations of various developmental stages and age categories. The results of the present study are comparable with other studies and also concludes The London Atlas as a reliable tool for age estimation. The London Atlas will need to be performed in different countries or ethnic groups to assess its robustness in age estimation across populations. One way of doing this would be to do a targeted study on Australian Aboriginals and Torres Strait Islanders versus Caucasoids. Reports have shown that Australian Aboriginals and Torres Strait Islanders have faster dental development than those of European descent [23]. A comparison of the London Atlas and another age estimation method on the specific target group would be beneficial in determining which method is more accurate. Furthermore, obtaining a large enough sample was a difficulty this study faced. This could be made possible in the future by utilising radiographs from multiple Queensland Health facilities.