Sixty years of research in dental age estimation: a bibliometric study

Abstract

Background

Dental age estimation (DAE) research has grown rapidly and became one of the biggest topics in forensic odontology. This study aimed to evaluate the DAE research trends over the span of 60 years using bibliometric analysis.

Methods

Sampling was performed in the Scopus database using a search string (“Dental Age Estimation” OR “Age Determination by Teeth”) to detect DAE-related studies. The search was performed from inception to the year 2022. A data-cleaning intervention using a fuzzy-matching technique was done to unify the author and affiliation name variations.

Results

The initial search returned 1638 articles, years of publication ranging from 1964 to 2022, with an approximate growth rate of 5.9% a year. Source analysis showed that most of the top article sources were Forensic Science International (n = 200). Cameriere R presents the overall highest score (77 articles, Local h-index 30). Authors from Shanghai Jiao Tong University produced the highest number of publications (n = 111). The most locally cited study was “A New System of Dental Age Assessment” by Demirjian et al. (Hum Biol 45:211-227, 1973) (n = 1507). The trending topics analysis shows that earlier DAE studies were focused on dental regressive changes and later changed focus to utilizing technological advancements. Institutions and Author's collaborations were also found to be internationally diverse with 20.82% of the articles being a product of international co-authorships.

Conclusions

DAE research has grown rapidly helped by multiple advancements in various technological ends. Along with the high demand for DAE analysis, authors and publishers need to continually improve their standards for their respective research and reporting and continue to increase collaboration.

Background

Dental age estimation (DAE) is one of the branches of forensic odontology, which aims to provide an estimation of an individual chronological age through an age-related dental variable (Greene et al. 2013). DAE methodologies require a firm quality assurance to reach a reliable and reproducible approach, starting from population sampling to the creation of models or tables. Therefore, it is essential that the forefront of DAE methodologists publish their results in peer-reviewed journals. The need to make scientific evidence available to the academic community is also corroborated by legal mechanisms. According to Daubert’s standards that rule the admissibility of evidence in Court in certain jurisdictions, methodologies used to produce evidence must be subjected to peer review and publication (among other requirements) (Lesciotto 2015). Moreover, Daubert’s standards also highlight the importance of knowing the method’s error rates before presenting forensic evidence (Fradella et al. 2003). This is especially relevant in the field of DAE because forensic casuistic has increased worldwide because of globalization (Thevissen et al. 2010b, 2009) and the influx of undocumented immigrants (Manica and Gorza 2019). This phenomenon creates a need for researchers to validate methodologies to calibrate models or identify error rates from other studies in their own population of interest (Bittencourt et al. 2018; Franco et al. 2021; Liversidge 2015; Merdietio Boedi et al. 2022). Consequently, DAE research has grown rapidly, with multiple original methods, validations, and modifications, and became one of the biggest topics in forensic odontology (Liu et al. 2016).

The trend of an expanding research field needs to be analysed to understand properly how much a particular research branch — in this case, DAE — has grown, particularly using bibliometric analysis. Bibliometric analysis is one of the approaches in scholarly research to understand emerging trends, collaboration patterns, and research gaps in scientific discoveries (Donthu et al. 2021). Although similar research has been conducted in forensic odontology (Liu et al. 2016; Sengupta et al. 2020), a focused bibliometric study in DAE has not been conducted so far, specifically a large-scale analysis of DAE publications. Therefore, the present study aimed to evaluate the DAE research trends over the span of 60 years using bibliometric analysis.

Methods

Data extraction

This is an observational descriptive study with retrospective data collection. Sampling was performed in the Scopus database to retrieve the metadata from peer-reviewed research in the field of DAE. The search string used to detect DAE-related studies was (“Dental Age Estimation” OR “Age Determination by Teeth”). Both terms are part of the common glossary used to index DAE-related studies in reference databases. The search was performed from inception to the year 2022. Only studies written in the English language were included, as determined by the filter set on the Scopus search prompt. The data was accessed and exported on January 1st, 2023. Data detected in Scopus was exported in.csv format to RStudio (version 3.4.0, R Foundation for Statistical Computing, Vienna, Austria).

Data cleaning

During the initial database export from Scopus, it was identified that multiple-character strings were not uniform due to variations in abbreviation, language, and characters. For example: (1) author “Różyło-Kalinowska, I” (Scopus ID 6603955520) was also listed as “Rózyło-Kalinowska, I”; (2) “Katholieke Universiteit Leuven” was also listed as “KU Leuven” or “Catholic University Leuven”; and (3) “Universidade de São Paulo” was also listed as “University of São Paulo” and “University of Sao Paulo”. Therefore, a data uniformization protocol was conducted for the two most important data points: author names and affiliations (Fig. 1). This process was accomplished by two reviewers (RMB and AF).

Fig. 1

Workflow diagram of the bibliometric analysis

Author names were standardized by cross-referencing the Scopus Author ID to the author names. If there were multiple authors with “similar” names coming from one Scopus Author ID, the names were joined together. However, the affiliations did not have the same unique number identifier as the authors recorded in the extracted bibliometric database. Therefore, a different approach was taken through fuzzy matching using Levenshtein Distance filtering (Levenshtein 1966). Fuzzy matching is a technique to identify a sequence of letters that are similar but not necessarily identical. It is often used to match data that has been entered into a database with inconsistent or incomplete information. Although fuzzy matching can work by various algorithms, Levenshtein distance was chosen in this research due to its filtering simplicity when compared to other edit distance measurements such as Jaro-Winkler or Hamming distance.

In the context of this study, Levenshtein distance was used to measure the similarity between different institution names. The affiliation name sequence was first converted to a number, and the shortest distance between the name was displayed. The threshold used for the Levenshtein distance filter was 0.1, and the filter results were joined together if a similar distance was detected for a particular naming sequence. This process was accomplished using the base R “agrep” function and tidyverse for filtering (Wickham et al. 2019).

Bibliometric analysis

The bibliometric analysis — including citation analysis, collaboration, and ranking — was performed using the Bibliometrix package (Aria and Cuccurullo 2017). It is important to note that all the citation counts and impact (i.e. h-index) were calculated locally, which means that it only considered the citations within the included data. The keyword trending topics were pooled via Bibliometrix with a word minimum frequency of 15 and three words per year. Additionally, a list of synonyms was created to combine similar keyword occurrences in the keyword trending topics. The collaboration of authors and institution network was visualized with Bibliometrix.

It is anticipated that despite the efforts to accurately represent the current research state of DAE by conducting data uniformization, some of the numbers presented in this study may not reflect the full status of the research topic due to the inherent complexities and nuances of the bibliometric dataset and analysis.

Results

The initial search returned 1638 articles, years of publication ranging from 1964 to 2022, with an approximate growth rate of 5.9% a year. The articles were written by 4357 authors, with an average of 4.11 authors per document. Almost 21% of the articles were a product of international co-authorships (Table 1).

Table 1 Overall information of dental age estimation studies

Source analysis

Source analysis showed that most of the top article sources were focused on forensic and anthropological sciences, with the top source being Forensic Science International (Elsevier, ISSN: 1872–6283, n = 200). The top source with the highest local h-index was also found to be Forensic Science International, with a local h-index of 49 and a total citation count of 7841 (Table 2).

Table 2 Study sources and local impact

Local impact

Among all the authors’ calculated impact, Cameriere R (Scopus ID: 6507826165) presents- the overall highest score with 77 articles and a local h-index of 30 (Table 3). In the affiliation’s productivity, authors affiliated with Shanghai Jiao Tong University (Affiliation ID: 60025084) produced the highest number of articles totalling 111 authored articles (Table 4). Considering the country’s scientific productivity, India (n = 667) is the country that produced more articles, but the United Kingdom is the country with the highest local citation count of 4440 (Table 5).

Table 3 Author’s productivity and local impact

Table 4 Author’s affiliation productivity counts

Table 5 Author’s country productivity and local impact

Article and keywords

Article analysis revealed that most of the top 10 studies in DAE research were a study of a new DAE methodology (Table 6). The most locally cited study was “A New System of Dental Age Assessment” by Demirjian et al. (1973) (Demirjian et al. 1973), with a total local citation of 1507. The second most locally cited study was “Brief communication: The London atlas of human tooth development and eruption” — a.k.a London Atlas — by AlQahtani et al. (2010) (AlQahtani et al. 2010) and followed by Lovejoy (1985) in “Dental wear in the Libben population: Its functional pattern and role in the determination of adult skeletal age at death” (Lovejoy 1985), which was also the only single-authored study in the top 10 articles.

Table 6 Top 10 most locally cited articles in dental age estimation research

Keywords occurrence analysis revealed that 23% of the studies were using “Forensic Odontology” as their main keyword (n = 537), followed by “Age Determination by Teeth” (n = 467), which was also the same MeSH term used for DAE research (Fig. 2). The trending topics analysis revealed that most of the DAE studies in the 1990s were focused on dental regressive changes. This trend then later evolved in the current days by utilizing more technological advancements, such as cone-beam computed tomography (CBCT) or machine learning (Fig. 3).

Fig. 2

Overall keyword count

Fig. 3

Trending topic analysis based on keyword year occurrences

Networks

Collaboration networks were visualized for citations, institutions, and authors. In the citations network, the majority of papers referenced “A New System of Dental Age Assessment” by Demirjian et al. (1973) as their study (Demirjian et al. 1973). Institution collaborations showed a diverse geographical location worldwide being the University of Macerata (Affiliation ID: 60027141), the most collaborative with other institutions from Europe (i.e. University of Zagreb [Affiliation ID: 60008408]), South America (i.e. Universidade de São Paulo [Affiliation ID: 60008088]), and Asia (i.e. Universitas Airlangga [Affiliation ID: 60069383]). Results for author collaborations were found to be internationally diverse — as seen in the 20.82% of international co-authorships, but there were small, isolated clusters from different authors from Chinese institutions. Even so, Liversidge HM (Scopus ID: 6602158831) was observed to be connected to every author cluster.

Discussion

The research in DAE has experienced tremendous growth in recent years and it can be attributed to several factors. Firstly, the Internet has made it easier for academics to access and communicate with their peers, enabling fast and effective coordination between researchers as demonstrated by the amount of international collaboration in our dataset (20.82%) (Finch et al. 2013). Secondly, technological advancements have opened new possibilities for data transfer and methodology in various studies. Data transfer was made easier through digital imaging formats (i.e. DICOM) and although original methodology validations are still common for different populations, researchers have adapted these methodologies with the latest advancements. For example, Demirjian’s staging has been adapted to include predictive modelling calculations (Galibourg et al. 2021) or image classification systems (Mohammad et al. 2021). Lastly, the demand for DAE reports in various identification matters, particularly in living individuals, has led to a continuous effort to support the legal systems worldwide, especially in protecting the rights of vulnerable age groups (i.e. children, pensioners) or geographic jurisdictions (i.e. asylum seekers) (Manica and Gorza 2019). These legal issues often require proper age assessment, and multiple DAE systems have been established to address this matter (Nuzzolese et al. 2011; Thevissen et al. 2010a; Zelic et al. 2016). These three factors have motivated the increase in DAE research and, as a result, increased the quality of the studies and moved the overall DAE studies to a more digital-based approach, utilizing mainly non-invasive approach through radiological data.

The top 10 sources in this field have a focus on not only forensic odontology, but mainly forensic medicine, anthropology, and human growth. These three fields are closely related to forensic odontology and DAE, as DAE is used to help identify individuals in forensics cases (Pretty and Sweet 2001), predict the age of archaeological sites in anthropology (Katzenberg et al. 2005), and serve as the basis for DAE methodology in children and juveniles in human growth studies. Although the human growth aspect did not contribute significantly to the topic of DAE methods for adults, which primarily rely on regressive dental changes, it remains the foundation of the most widely used DAE methods, such as staging (Gunst et al. 2003), atlas (AlQahtani et al. 2014), or metric measurements (Cameriere et al. 2006).

In terms of author productivity and local impact, Cameriere R (Author ID: 6507826165) leads in all three aspects: article count, h-index, and citation count. Cameriere R is the founder of the AgEstimation project and the author of the well-known third-molar maturity index method (Cameriere et al. 2008), which has been used and recalibrated in various populations with satisfactory results. A study comparing its performance has been conducted and it is recommended that, even though the methodology can work for various populations and purposes (i.e. predictive or probabilistic modelling), a recalibration is necessary for the method to maintain its accuracy (De Micco et al. 2021). His methodology has had a significant impact on university rankings and trending topics. The University of Macerata (Università degli Studi di Macerata, Affiliation ID: 60027141) — the affiliation of Cameriere R — is ranked second in affiliation productivity (Table 4), and the trending topics analysis shows that the term “Third-Molar Maturity Index” has been a trend from approximately 2018 to 2020 (Fig. 3).

The first ranking in affiliation productivity is Shanghai Jiao Tong University (Affiliation ID: 60025084), with the starting initial recorded research in 2013 by Yanet al. (2013) which studies the usability of Demirjian’s method in multiple populations through meta-analysis. It was suggested that there is a need for population-specific formula recalibration of this methodology (Yan et al. 2013). Additionally, the top locally cited article from Shanghai Jiao Tong University is an article by Ye et al. (2014) (n = 40), which studies Demirjian’s method and Willem’s formula in the Chinese children population (Ye et al. 2014).

The top 10 most locally cited articles give a glimpse of how wide the DAE research scope is. Although most of the articles focused on the modern population, there are also articles that used DAE in anthropology. The most locally cited article, “A New System of Dental Age Assessment” by Demirjian et al. (1973) was seen as a foundation of every modern DAE study (Demirjian et al. 1973). This can be observed in Fig. 4, that the most well-known article in DAE uses Demirjian staging research as their reference. As explained before, the Demirjian staging method has been validated into multiple populations and adaptations using a modern statistical approach, and other staging modifications. Additionally, most of the top articles focused on a non-invasive approach to obtain an estimated dental age as the usage of these methods is not limited to the deceased, but also to the living population.

Fig. 4

Local co-citation network of dental age estimation studies. The naming follows the surname of the first author and followed by the publication year of the research

The connection between the authors and the institution is reflected in Figs. 5 and 6. The connection can be seen as a cluster of collaboration, either by similar study interest, organization, generational, or geographical language aspect. These aspects may create or separate the connection between researchers, as authors or institutions can have a specific interest in a certain subject or approach (i.e. legal age, third molar development, specific modalities). Additionally, most of the authors and institutions are connected through forms of an international organization, such as the International Organization for Forensic Odonto-Stomatology (IOFOS). IOFOS is an organization for forensic odontology expert to meet and refine the standardization of forensic odontology methodology, including DAE. Another prominent organization is “Arbeitsgemeinschaft für Forensische Altersdiagnostik” (AGFAD) led by Schmeling A. This organization gives an overall insight into age estimation in general and not necessarily in DAE. Even so, reports and standardization that come from AGFAD have played a major role in DAE research (Schmeling et al. 2008). This global research environment means that even though there are known variations of human dental changes throughout the population worldwide (Liversidge 2008), most of the research questions are shared through collaborative research. As stated by Espinoza et al. (2022), these conditions give a very powerful incentive for methodology development through a form of international collaborations and multicentre-level studies (Espinoza-Silva et al. 2022).

Fig. 5

Institutional collaboration network in dental age estimation studies

Fig. 6

Author collaboration network in dental age estimation studies

The limitation of this study has resulted from the limitation of the academic database itself, or more accurately, the lack of standardization in data input. In this study, Scopus was chosen as the preferred database since Scopus offered a high-quality curated academic database to be used in a large-scale evaluative study (Baas et al. 2020). However, we observe multiple errors — in the author’s name or affiliation — in regard to the extracted database, which comes from a non-ASCII character and language. In the author’s name, it is considerably easier to unify the name due to the Scopus Author ID numbering. In terms of affiliations, Scopus also assign affiliation to a similar unique ID. Unfortunately, this information is not extracted from the data export, creating difficulties in data uniformization. We recommend that every bibliometric study runs a data cleaning to their database, since failing to unify the data may result in a misrepresented current research condition. Furthermore, we noticed that early research has an inadequate filing system, which resulted in (1) the research itself not being exported/detected and (2) incomplete bibliometric analysis. Due to this limitation, certain research might not have been adequately catalogued in the database we exported for our study. As a result, the absence of earlier studies in our dataset may have gone undetected and created an impression that research on DAE began in 1964, which is not the case. One example of undetected research is Gustafson’s influential work on adult DAE: “Age Determination of Teeth” (Gustafson 1950). The research itself is represented in our database as “cited research” in Fig. 4 but not in the overall database. This can be explained by the keyword index in Scopus for Gustafson’s work that only used “Age Factors” and “Tooth” when related to the MeSH indexing system. Consequently, these partially indexed early studies made the trending topic analysis — which uses keywords — partially missing, with only studies from 1989 onward can be analysed (Fig. 3).

Conclusions

DAE research trend has grown rapidly throughout the years. Together with advancements in various technological ends — such as advanced imaging technique and analysis, DAE research has evolved to a more robust methodology through validations and focused more on non-invasive techniques. The highly collaborative environment in DAE research shows that even though DAE methods were mostly population-specific, significant effort has been made between research centres to solve the variability problem, either by validation or the creation of a new DAE approach. Along with the high demand for DAE analysis, authors and publishers need to continually improve their standards for their respective research and reporting and continue to increase collaboration.