Third molar maturity index cut-off value optimization for legal age of majority in Indonesian juveniles

INTRODUCTION

The knowledge of an individual’s chronological age (CA) is an important aspect in legal matters, especially in determining whether an individual has reached adulthood or not. When the information of CA is not available, age assessment is required. In terms of dental- related age assessment, the specification of third molar is widely encouraged to determine CA threshold of legal interest [1]. Multiple methods have emerged, from the utilization of dental staging [2] to metric measurement of open apices. The latter can be done by assessing the third molar maturity index (I3_M) proposed by Cameriere et al. [3], and this approach has been tested throughout multiple populations with satisfactory results [4].
Nevertheless, the capability of I_3M to accurately predict adulthood of juveniles relies heavily on population-specific data references, which gives an output of a different I_3M cut-off value of each population and sex [5]. The main cut-off value proposed by Cameriere et al. [3] is 0.08, with an I_3M value below 0.08, indicating an individual of a higher probability of adulthood, and vice versa. However, a Brazilian sub-population study conducted by Goetten et al. [6] reported that in northern Brazilian males, the original cut-off value can be used, but the value needs to be re-calibrated to 0.12 for females, in order to reach the optimal model per-formance. This evidence emphasized the importance of country-specific adaptations to achieve high classification performance of the I_3M cut-off value [7].
In Indonesia, the legal age is 18 years. However, a recent pilot study by Boedi et al. [8] reported that the I_3M cut-off value for legal age of marriage (which is 19 years) was 0.08 for males and 0.09 for females. This finding indicates that the development of third molars in Indonesian juveniles was slower; therefore, a cut-off value re-calibration to optimize its performance is needed.

OBJECTIVES

The aim of this study was to assess the performance of I_3M in Indonesian population as well as the calculation of a new cut-off value in Indonesian juveniles to determine whether an individual has reached age of 18 years.

MATERIAL AND METHODS

The sample consisted of 900 panoramic radiographs, with an even distribution of chronological age and sex, from 15 to 23.98 years (mean CA, 19.49 ± 2.61 years; Table 1). All samples were taken from Pramita Laboratory, Semarang, Central Java, Indonesia. The cohort consisted of Javanese origin residents from Semarang, Central Java, with middle to upper-middle categories of incomes. The sample was selected based on the presence of lower left or lower right third molars, without any visible radiographic abnormalities or dental treatment. Sample data anonymity was maintained throughout the study, with only information on sex, date of birth, and date of exposure recorded. Ethical approval was obtained from the institute’s ethics committee (approval No. 139/EC/KEPK/FK-UNDIP/VI/2020).
Panoramic radiographs in .jpg format were measured using GIMP version 2.10 built-in tools (GIMP Deve­lopment Team, 2024). The measurement was done with lower left third molar being the primary observation. If the lower left third molar was not present, lower right third molar was measured. The I_3M measurement was performed as described by Cameriere et al. [3], where the sum of lengths from inner margins of the tooth apical ends was divided by the tooth length from the apical to the end of the highest point of the crown. If apices were closed, I_3M was counted as 0. I_3M measurements were done by two observers, i.e., JRK and RMB. JRK is an oral and maxillofacial radiologist with over 4 years of experience, and RMB is a forensic odonto­logist with over 8 years of experience in dental age estimation. For this purpose, 50 randomly selected panoramic radiograph images from the sample were used for inter- and intraobserver tests. All measurements were performed with observers blinded to chronological age and sex of each image.
Measurements were recorded in an Excel file and processed using R software (R Foundation for Statistical Computing, version 4.0.5). Inter- and intra-observer measurements were done with intra-class correlation coefficient (ICC). The analysis was performed using logistic regression to calculate the probability of an individual being above or below 18 years old, given a certain I_3M value. Furthermore, the optimal cut-off value was calculated with a cut-point for I_3M to determine the legal age of samples. The model’s predictive accuracy was assessed using receiver operating curve (ROC) and area under curve (AUC). Additionally, the optimal cut-off value was determined with the highest Youden’s index value.

RESULTS

The intra- and inter-observer reliabilities rates were 0.92 and 0.85, respectively, indicating an excellent agreement of measurements between the observers. The initial t-test result provided a significant difference between males and females I_3M (p < 0.05) (Table 2), with a correlation coefficient value of 0.66. The probability of an individual being above 18 years old, given a certain I3_M, is described in Table 3 for males and females. It was observed that the probability of age above 18 years in female samples had a steeper curve, meaning that the utilization of I_3M was less sensitive (Figure 1).
The employment of sex-specific cut-off value can achieve a better accuracy result for males, with a non-sex-specific cut-off value providing an acceptable accuracy for both sexes. Cut-off value optimization for both males and females resulted in 0.1, with an accuracy level of 0.88, meaning that if an individual has an I_3M ≤ 0.1, there is a higher probability that the person’s age is above 18 years. In sex-specific cut-off value, I_3M ≤ 0.13 with 0.85 accuracy was observed in females, and I_3M ≤ 0.1 with 0.93 accuracy in males. The full performance results are described in Table 4.

DISCUSSION

In Indonesian law, the legal age of majority is 18 years. This means that once an individual reaches 18 years of age, he/she is treated as an adult in legal matters, bearing full responsibility of their involvement in civil trials. However, when dealing with minors, Indonesian law suggests a se­parate approach focused on re-education and rehabilitation. Therefore, determining whether an individual’s age is above or below 18 years is crucial in court proceedings nationally or internationally, particularly when a person lacks legal identification [9]. Nearly half of the children in Indonesia do not possess legal documents, especially in rural are-as [10]. This issue can significantly complicate trials, in which age estimation can provide scientific evidence with probability and observation of individual growth.
Multiple age estimation methods can be employed to determine a person’s legal age, such as skeletal me­thods and dental radiographs. Generally, it is advised that the amount of radiation exposure to an individual should be kept to a minimum [11]. Alternatively, to eliminate radiation exposure, age estimation through magnetic resonance imaging can also be utilized [12]. In dental age estimation, the specification of third molar in legal age research is mostly done with two distinct approaches, such as staging [13] and metric measurements [14]. The American Board of Forensic Odontology recommends that when analyzing an individual’s age, two or more methods with a similar population sample must be used [7]. However, these two methods should not be combined to achieve a single result without a proper statistical weighting or analysis [15].
The I_3M method is a popular approach used in dental age classification models, and has been adopted worldwide in different populations. However, calibration to each specific population and adaptation to local governing laws are necessary [4, 6]. For example, Balla et al. [16, 17] utilized the I_3M method to protect children in south Indian population against child labor, and assessed the minimum age of criminal responsibility. Whereas Angelakopoulos et al. [18] in their multi-ethnic I_3M studies reported that there is a need of cut-off value calibration for each country, as the variability in the accuracy was significant.
The biological timing of third molars’ growth is influenced by multiple factors, including genetics and environment. Therefore, the geographical distribution of dental maturity and physical development vary widely for third molars, and there is a need for population validation for each methodology in dental age estimation [19]. This validation of cut-off values can be achieved through two approaches, including testing the performance of a similar population’s cut-off values, or their re-calibration. The original cut-off value of 0.08 was tested in various population with satisfactory results; however, several studies have reported a need of a re-calibration [6, 20-23] (Table 5). Goetten et al. [6] recommended I_3M ≤ 0.12 for northern Brazilian females, and Balla et al. [20] proposed I_3M ≤ 0.17 for south Indian population with impacted third molars. Another extensive study by Angelakopoulos et al. [18] reported various re-calibration values for multiple populations. Nevertheless, if testing reveals poor accuracy or may achieve a better performance, re-calibration of cut-off values becomes necessary, even in a sub-population of a country [23].
This population-specific cut-off re-calibration was demonstrated in the current study among Indonesian population. A previous research that utilized the I_3M method for determining the legal age of marriage in Indonesia (distinguishing individuals’ age above or below 19 years), established cut-off values of 0.08 for males and 0.09 for females [8]. These results indicate that the third molars’ growth is slower in the current population, as the I_3M value for 19 years old individuals in Indonesia and 18 years old ones in other countries are comparable [24, 25]. However, several populations present a similar rate of growth of I_3M values, as reported in Table 5. Furthermore, two methods of determining the legal age of Indonesian juveniles were proposed, such as using I_3M ranges and cut-off values for each sex. The utilization of I_3M range is straightforward, and involves observation of the mean probability of legal age, given a certain range of I_3M values. It was observed that if an individual has an I_3M < 0.04, his/hers probability of being over 18 years old is above 0.72 for males and 0.92 for females, respectively. Regarding cut-off values, three values were calculated, i.e., non-sex specific, male-specific, and female- specific. However, given that the accuracy of female-specific cut-off value is lower compared with non-sex specific, a universal cut-off value of I_3M ≤ 0.1 is therefore encouraged.
The current study has several limitations, including the lack of methodological comparison to determine legal age and the scope of the sampled population. Alternative methods for determining legal age, such as staging and calculating its probability, were not explored [26]. Through staging, a combination of multiple third molars can be utilized [2]. However, the risk of multi-collinearity arising from the parallel growth of third molars needs to be considered [13]. Additionally, in terms of population sampling, representation of multiple populations and ethnic groups is necessary, as Indonesia covers various ethnic groups, which spread across the nation, and this fact may affect differences in third molars’ growth [27]. Future studies should aim to address these limitations by testing the performance of staging-based legal age probability calculation, and evaluating the current cut-off values across different ethnic groups in Indonesia.

CONCLUSIONS

The third molar maturity index can be an effective tool for accurately classifying whether an Indonesian individual is above or below 18 years of age. It is emphasized that the cut-off value re-calibration is needed in Indonesian juveniles, with a recommended cut-off value of I_3M ≤ 0.1, along with the inclusion of mean probability of a certain I_3M range. Despite variations in optimal cut-off values between males and females, the utilization of a non-sex-specific cut-off value is recommended for practical purposes.

DISCLOSURES

1. Institutional review board statement: The study was approved by the Ethics Committee of the Universitas Diponegoro, Indonesia, with approval number: 139/EC/KEPK/FK-UNDIP/VI/2020.
2. Assistance with the article: None.
3 Financial support and sponsorship: None.
4. Conflicts of interest: The authors declare no potential conflicts of interest concerning the research, authorship, and/or publication of this article.

References