The reliability of tooth development as a human biological growth marker stems from its resistance to influences such as genetics, nutrition, climate, hormones, and the environment. Being largely independent of exogenous factors and seldom affected by pathological conditions further establishes it as a dependable growth indicator in human biology [23,24,25]. Age estimation, particularly in young adults, primarily relies on the morphological characteristics of bones and the mineralisation of third molars. After age 14, third molars are the sole teeth undergoing developmental processes, making them crucial for forensic age estimation [26, 27]. Despite variations in position, morphology, and development, these teeth remain a focal point, especially for distinguishing between adults and minors. Existing literature supports extensive research on these teeth concerning their development, emergence, calcification, and regressive changes linked to chronological age [10,11,12,13,14]. However, this study, to the best of the authors’ knowledge, represents the first attempt to assess the capability of Robert et al.‘s stages of relative widths of root canals in lower molars [17] for discriminating south Indian juveniles and subadults concerning the age threshold of 18 years.

Robert et al. suggest that the pattern observed in each category of root canal width is founded on the premise that as individuals age increases, the relative mesiodistal widths in the distal roots of all lower three molars decrease [17]. In our study, we noted a trend of root canal narrowing following the closure of the apex in the molars. The findings from this validation study indicated that all males and females in the studied population with stage H left molars and RCW-C pattern were over 18 years old. These results align with the findings reported in the original study [17] and two validated studies conducted on South African and Maltese populations [18, 19]. In males, the most common patterns were RCW-A and RCW-C, each accounting for 40%, while in females, RCW-C was the most prevalent at 42%, followed by RCW-B at 33%. These findings contrast with Tangkabutra et al.‘s report, [19] where they identified RCW-C as the least common pattern. Additionally, the results demonstrated that as the categories increased (RCW-A to C), there was a corresponding rise in the mean chronological age for both males and females. Furthermore, the first (25th percentile), median (50th percentile), and third (75th percentile) quartiles displayed an increase in age from RCW-A to RCW-C. This suggests no significant differences between sexes in RCW development when a specific pattern was assigned.

A Spearman correlation test was used to evaluate the strength and direction of the linear relationship between RCW patterns and chronological age. The results showed a moderate positive correlation for both males and females, meaning that RCW patterns tend to increase with age. However, the moderate correlation is not easily explained. It is likely that the limited number of RCW patterns in the classification does not adequately capture the chronological progression of the characteristic being studied. Various factors influence the accuracy of age estimates, and one crucial aspect is the observer’s performance. The precision of dental age estimation is closely tied to the assessment method employed [28]. Despite the well-established characteristics distinguishing each developmental stage, the classification of tooth developmental stages can be subjective and susceptible to inter- and intra-observer variability [29]. Our study found a satisfactory intra-observer agreement (0.75), but the inter-observer agreement was comparatively lower (0.67). These results exhibited lower inter- and intra-observer agreements compared to the original study, where agreements surpassed a kappa score of 0.9 [17]. However, they align more closely with the outcomes of Davidson et al. [18] and Tangkabutra et al. [19]. The differences observed compared to the original study may be attributed to the greater experience of the original method developers [19]. Similar observations were noted in a study by Timme et al., [20] where they reported poor agreement between observers, i.e., 0.07 (95% CI: −0.11–0.26; 40.7% agreement) for females and 0.11 (95% CI: −0.08 − 0.31; 41.7% agreement) for males, indicating limited consistency. They suggested that the lack of reference points in the method might contribute to poor agreement between observers. Recognising patterns becomes challenging, especially considering the variations in shape and size between the first and second molars compared to the third molars. They further emphasised the need to consider this pattern recognition approach carefully, particularly in light of inter-observer agreement concerns.

In forensic age assessment, experts frequently need to determine the most likely age and/or the minimum age of the individual being examined. In 2016, Schmeling et al. [30] introduced the “minimum age concept,” aimed at preventing the misclassification of minors as adults. This concept is designed to provide a high level of certainty regarding the attainment of the age of majority. However, it often relies on a single observation rather than considering the entire reference sample centered around the mean age of the stage. In this study, we incorporated the minimum age concept to mitigate the risk of age mimicry bias. Our findings suggest that stage RCW-C indicates a minimum age of 18.73 years in males and 19.01 years in females, respectively. These findings are consistent with those of the original study, where the minimum age for RCW-C was reported as 18.16 years for males and 18.45 years for females [17]. Additionally, our results align with the conclusions drawn by Davidson et al. and Tangkabutra et al., emphasising the persistent capability of RCW-C in effectively differentiating between adults and minors [18, 19].

Legal professionals and governmental entities frequently seek the expertise of forensic specialists for age estimation in individuals involved in civil and criminal litigations. However, the necessary probabilities to meet the criteria for age estimation differ depending on the context of the assessment. The requirement is typically at least 90% for criminal proceedings, whereas it is 51% for civil proceedings [30]. In criminal law, there is a stringent demand for age assessment or classification to be provided with a very high probability, with a particular emphasis on minimising false positive rates [31]. Garamendi et al. have distinguished errors in age assessment into two categories: technically unacceptable errors (false negative) and ethically unacceptable errors (false positive). Ethically unacceptable errors, especially in criminal proceedings, are considered the most harmful to the child’s best interest [32]. Therefore, it is crucial to thoroughly investigate and accurately report the false positive rate associated with the applied methods. In this context, stage RCW-C in our current study achieved a specificity of 100% in both males and females. However, our findings indicate that RCW-C exhibits lower sensitivity, suggesting a potential for a high percentage of adults being incorrectly classified as minors in the studied population. Therefore, the conclusion drawn from these findings is that when an individual, whether male or female, presents with fully formed mandibular first, second, and third molars (stage H of Demirjian’s staging system) for age assessment, and stage RCW-C is concurrently observed, there is a strong likelihood that the individual has surpassed the legally relevant age of 18 years. However, it is essential to exercise caution in its application due to inherent limitations.

There are several methodological issues in identifying root canal width (RCW) patterns. The first issue concerns root canal morphology. The initial assessment involves making decisions about the distal root canals, specifically the mesiodistal width of the root pulp in all lower three molars. Singh et al. [33] examined the root canal morphology of the lower three molars in South Indians, finding that 90% of the lower first molars and 91% of the lower second molars were two-rooted, while only 63% of the lower third molars had two roots. This variability might limit the method’s applicability to the studied population. Another issue relates to the original RCW classification. The foundational work by Roberts GJ et al. [17] proposed a distinct classification of RCW stages, illustrated in Fig. 1. However, our systematic analysis revealed discrepancies between these patterns and those depicted by Roberts GJ et al. [17] Subsequently, Davidson et al. [18] introduced an additional category, RCW-U, to accommodate patterns not fitting into the initial three categories. Studies have reported that the prevalence of the RCW-U category ranges from 15.47 to 35.87% [18, 19]. In our study, we found that less than 10% of the samples fell into the RCW-U category. The most common RCW-U pattern we observed was LL6 < LL7 = LL8 (73.8%), which aligns with Davidson et al.‘s finding of 71.43% for this pattern. The reasons behind these varying patterns are unclear but might be due to differences in secondary dentin deposition rates among individuals or variations in panoramic radiographic acquisition, projection geometry, and the focal trough, all of which can affect the image characteristics of the root canal width in third molars [17]. This study has several limitations. Firstly, it focused on a somewhat geographically restricted sample, comprising participants from five South Indian states and excluding other regions in India. Future research could benefit from validating this method in diverse geographic areas, establishing a more representative and comprehensive sample. Secondly, the method’s reliance on the presence of all three molars poses a challenge, given the notably higher prevalence of congenitally missing lower third molars in the studied population [34, 35]. Additionally, the evaluation of the method may be constrained in samples lacking a clear definition of root canals, particularly in impacted lower third molars where visibility is hindered by physiological thickening and compaction of the mandibular bone in the wisdom teeth region [20]. Factors such as agenesis and impaction of third molars could influence the method’s application or the assignment of the RCW pattern, negating its use as a maturity marker. Thirdly, another limitation is the potential for radiographic distortion or alterations in the projection of root canal width onto the film or sensor due to the use of two-dimensional images [36]. The use of cone-beam computed tomography (CBCT) could mitigate this issue, providing more accurate measurements.

Despite these limitations, which are common to many age estimation methods, we believe this approach paves the way for exploring new methods to determine the legal age threshold of 18 years. Additional research employing this methodology is needed across varied populations to evaluate its effectiveness and to compare the findings of this study, thereby strengthening confidence in this approach.