Experts’ opinion on the importance of therapeutic features for dental human identification using intraoral radiographs

Odontology is one of the primary means for human identification. As common components of dental practice, intraoral radiographs represent an important source of antemortem (AM) data. Interpretation of these radio-graphs, however, may vary between forensic odontologists. The aim of this study was to investigate experts ’ opinions on the interpretation of dental identifiers from intraoral radiographs. A web-based survey was distributed to international forensic odontologists. The survey included simulated cases of human identification and questions related to the use of intraoral radiographs in dental identification. Each case required the visual radiographic comparison of one AM and one postmortem (PM) intraoral radiograph. Further questions were related to the importance of different parameters that can be examined through intraoral radiographs, such as morphological, therapeutic, and pathological features; while final questions were asked about the methods used for comparative dental analysis. The survey was answered by 57 forensic odontologists across the world. The simulated case that involved therapeutic features (dental restorations) was positively identified by 56 participants (98%), while the case without therapeutic identifiers was identified only by 14 participants (24%). Almost half (54%) of the participants mentioned therapeutic features as the best (when available) for comparative human identification, followed by morphological features. The use of comparative methods for radiographic analysis was mentioned by 18 participants (32%). Intraoral radiographs enabled the analysis of dental and non-dental features – that can be therapeutic, morphological, and pathological. Specifically, therapeutic features allow a more straightforward comparison, while morphological comparisons require detailed visualization, and knowledge of anatomic structures and their variations.


Introduction
The comparative dental analyses figure as primary means for human identification [1,2].In practice, primary means are considered scientifically reliable and stand-alone toolsthat do not necessarily require combination with other methods, such as friction ridge and genetic analyses.In the comparative dental analysis, antemortem (AM) data obtained from dental records are compared with postmortem (PM) data obtained from the dental autopsy.The dental features that are analyzed in AM and PM data can be therapeutic, morphological, and pathological [3].Radiographs play an important part in this process because they provide a visual representation of the victim's dental status.Their importance is highlighted by the guidelines [4] for human identification of the American Board of Forensic Odontology (ABFO), in which the radiographic assessment of the victim is encouraged despite the condition of the cadaver (e.g.preserved, decomposed, charred).Moreover, authors have demonstrated that radiographs are so valuable that they can change the conclusion of expert's report [5].
In this context, two main problems emerge: 1) the quality of available AM radiographic evidence [6,7] and 2) the subjective interpretation of the AM/PM radiographic evidence made by forensic odontologists.When it comes to the quality of AM data, forensic odontological reports depend on what is available.Dental records partially filled and with incomplete radiographic registration of dental status are common [8][9][10].Examples of valuable partial information from the victim's dental status come from intraoral radiographs.These radiographs are frequently obtained for endodontic treatment, restorative dentistry and periodical check-up.Hence, they are potential sources of AM evidence in form of therapeutic features.As any other type of radiograph, they also register morphological and pathological features that may be useful for human identification [10][11][12].Interpreting images and detecting dental identifiers, however, are subjective decision-making steps during the analysis and comparison of AM/PM data.The effect of these steps in the reconciliation phase may depend on the way that forensic odontologists visualize and weight therapeutic, morphological and pathological dental features.
In order to understand a little more forensic odontologists' minds, this study aimed to propose a survey to assess the opinion of forensic odontologists regarding simulated cases of AM/PM data comparison using intraoral radiographs.

Materials and methods
This study has been reviewed and approved by the University of Dundee (UoD), School of Health Sciences and Dentistry Research Ethics Committee (SREC), application number UOD-SDEN-2021-006.
The study consisted of a survey that included three simulated cases of human identification using intraoral radiographs and AM and PM data.The survey was structured and conducted via the Joint Information Systems Committee (JISC) v.2021 software.The survey targeted international forensic odontologists members of the Association Forensic odontology for Human Rights (AFOHR) and the British Association of Forensic Odontology (BAFO) via email and end-to-end messenger (WhatsApp).The latter was addressed through a private group in which access is granted by means of invitation only.
The survey was designed with different categories ("working life details", "casework-based response", and "personal experience during the reconciliation process") of questions, in which 12 questions (seven open-ended and five multiple-choice) were organized.The first category and consisted of three questions that aimed to understand the demographics of the target sample.The second category consisted of "casework-based responses", in which three simulated scenarios with AM and PM radiographic data were presented to the experts.Scenarios #1 and 2 included radiographs that had therapeutic features (Figs. 1 and 2), while scenario #3 did not include evidence of dental treatment (Fig. 3).The experts were not informed about the fact that AM/PM images matched.Each of the scenarios required a comparative analysis of the images and their personal opinions (conclusion) about the cases.
The conclusions were classified as true positive (in case of positive identifications), false negative (in case the expert concluded that the case could not be identified), and other (giving to the expert the opportunity to explain, e.g.insufficient data).The third category was about experts' "Personal experience during the reconciliation process" and consisted of five questions (Table 1).
The outcomes of the survey were quantified and tabulated for descriptive statistics of absolute (n) and relative frequencies (%).For quantitative purposes, the answers regarding identification were categorized into "established", "probable", "possible", "not established" and "insufficient information".

Working life details
The survey was responded by 57 participants out of 101 -representing a response rate of 56%.Ten participants were from the United Kingdom, ten were from India, nine were from the United States of America, five were from Brazil, three were from Canada, three from Indonesia, three from Malaysia, two from Australia, two from Pakistan, and one representative for Austria, Belgium, Croatia, Italy, Netherlands, New Zeland, Nigeria, Norway, Poland and the United Arabic Emirates.
Most of the forensic odontologists (n = 27, 47%) has less than five years of experience, followed by forensic odontologists (35%) that had more than 20 years of experience.The remaining participants had between 5 and 20 years of experience in practice.Most of the participants (43%) were trained at master level, and seven (12%) held a doctoral diploma.

Casework-based response
Casework simulation #1 showed 98% of true positive responses (participants correctly associated the AM and PM intraoral radiographs).Only one participant did not consider the case a positive identification (Table 2).In casework simulation #2, nearly 25% of the participants reached a true positive response.Most of the participants (53%) considered the case with insufficient information to enable a more conclusive response.The remaining participants (29%) responded that both radiographs (AM/PM) did not belong to the same person (Table 2).In casework simulation #3, 51% of the participants responded with a true positive comparison.Oppositely, 21% of the participants answered that both radiographs did not belong to the same individual (false negative).The remaining participants would need more information to conclude about the case (Table 2).
In casework simulation #1, most of the participants (52%) relied on therapeutic features, while 39% considered morphological features as the main components to be considered for the comparative dental analysis in human identification.Casework #2 did not include therapeutic features, hence most of the participants relied on morphological features (43%) and the number and position of teeth (33%) for human identification.Despite the presence of restorative identifiers in casework #3, most of the participants mentioned morphological features (41%) followed by therapeutic features (23%) as the main identifiers during the comparative dental identification (Table 3).

Personal experience during the reconciliation process
In open-ended questions, the participants mentioned several dental features that were considered during the examination of radiographs for comparative dental identification.The features include those related to tooth pattern (e.g.present and missing teeth), morphology (of tooth parts), and other structures (e.g.bone-related) (Table 4).
When asked about the features considered the most important in routine human identifications, the experts selected a sequence of features (in order of importance) among the following: a) morphological > therapeutic > bone features; b) therapeutic > morphological > bone features; c) bone > morphological > therapeutic features; and d) therapeutic > bone > morphological features.The common sequence was "b" (n = 34, 54%), followed by "a" (n = 16, 26%).The other sequences were selected six times each.
When asked about the type of radiograph considered more valuable/ beneficial for dental human identification, 77% of the participants agreed on the equal contribution of periapical and bitewing radiographs.Out of the remaining participants, two preferred bitewings and 12 preferred periapical radiographs.

Discussion
Intraoral radiographs have been common sources of AM evidence [13,14], and the scientific literature has proposed standard procedures for the radiographic analysis of dental evidence in forensic sciences [15][16][17][18].Dental evidence can establish human identification even if the AM and PM data have only a few (but strong) similaritiesif this is the case [19].Judging the dental similarities as strong enough for positive identification, however, is a subjective decision.This study proposed a survey with case-based simulations to understand how forensic odontologists work in practice.
Therapeutic dental features have an important role in dental human identification, and proof of it is the detailed list of restorations and restorative materials included in the coding system of KMD PlassData DVI™ (INTERPOL's compatible solution for reconciliation in disaster victim identification) [20].The present study confirmed the importance of restorations with the outcomes of responses to casework #1, in which 52% of the participants based their opinion on the pattern and outline of restorations.For this casework, in particular, 98% of true positive answers were detected.Other studies in the field also addressed the distinctive role of restorative materials seen on radiographs for human identification [21][22][23].Borrman et al. [21], specifically, found divergent outcomes.The authors observed low level of accuracy among forensic experts during the reconciliation process between AM and PM dental evidence.The differences between outcomes can be explained by the different training of experts sampled in Borrman's study and our study.While the former sampled experts from several fields, namely radiology, odontology and general forensic sciences, our study sampled forensic odontologists only.Since dental human identification is a specialized task, investigating forensic odontologists (exclusively) seems to be the proper approach to understand the decision-making process in the radiographic reconciliation of dental evidence.
In comparison with the scientific literature, our study had outcomes that were more similar to those observed by Mclean et al. [22].According to the authors, the simulated reconciliation using bitewing radiographs led to an accuracy rate of 93%.Evidence of therapeutic features were used by the authors (as it was in our study), and led to high accuracy rates as we observed in our findings as well.In our study, however, we used bitewing radiographs specifically in casework #2, in which therapeutic features were not presented.In casework #2, tooth morphology was considered the main feature for identification (43%), and we found the lowest rates of true positive answers (24%).These outcomes suggest that morphological features are indeed fundamental for dental human identification, but the correct interpretation of these experience as a forensic dentist?Q4.Scenario 1: AM and PM are both periapical radiographs of the 2nd lower molar of the right side with a dental restoration (Fig. 1).
Casework-based response Q5.Scenario 2: Both radiographs show permanent teeth with no therapeutic treatments.AM is bitewing and PM periapical of upper posterior teeth of the left side (Fig. 2).Q6.Scenario 3: AM and PM are both radiographs of posterior teeth, in AM only a small occlusal restoration in 2nd premolar while in PM endodontic treatment is present in 2nd premolar (Fig. 3).Q7.The following questions were asked for each the three case scenarios above: -In your opinion, do the AM and PM radiographs belong to same individual?-What main features of the dental radiographs you analysed for comparison?Q8.In your opinion, which sequence is recommended on the bases of importance while examining intraoral radiographs (1 as the most important and 3 as the least important)?a) Tooth morphology; b) Dental treatment; and c) Bone Personal experience during the reconciliation process Q9.How confident would you be by assessing an intra oral radiograph WITH dental treatment for human identification?Q10.How confident would you be by assessing an intraoral radiograph WITHOUT dental treatment for human identification?Q11.Do you consider any specific method when comparing intraoral radiographs during comparative dental analysis for human identification?Q12.Which type of intraoral radiograph is more reliable while comparing AM and PM dental records?
AM: antemortem; PM: postmortem.features can be challenging and misleading.The subjectivity behind the analysis of morphological features was previously addressed in the scientific literature [23].Authors have demonstrated that experts might prefer to assess the morphology if roots instead of crowns [15].Additionally, the literature has indicated the preference of experts for the morphology of the roots of first molars and the shape of maxillary sinuses for human identification [23].These features, however, are not fully registered in bitewing radiographs.For proper visualization, periapical radiographs might be needed.
A more balanced distribution of dental features was observed in casework #3, in which forensic odontologists used a broad variety of therapeutic, morphological and pathological evidence to reach their conclusions.Interestingly, the radiographs presented AM and PM were obtained in different moments from the same individual.Consequently, the AM image did not show the obturation material used for root canal treatment neither the crown filling.This case illustrates an explainable discrepancy.The false negatives (21%) detected in this case could be associated with the explainable discrepancies.Additional evidence used to support expert's conclusions in this case consisted of tooth (41%) and periodontal morphology (21%).Specifically, the intraoral AM and PM radiographs used in the case clearly show the mental foramen and distinctive trabeculae of the mandible.Despite the presence of these features, it is clear from experts' responses that therapeutic features make them more confident and more accurate in the process of dental human identification.
From a technical aspect, the use of radiographs in this study was relevant to understand how forensic odontologists behave in a simulated scenario of human identification practice.Radiographs represent a common source of AM data [24], and in the PM context they are corroborated as fundamental means to register dental features [25].According to the Approved American National Standard's best practice recommendation 108, dental radiographs should be collected in case of disaster victim identification [25].Specifically, the Revised American Dental Association's technical report 1088 highlights that detailed and complete dental examination (via autopsy) is not feasible if a radiographic registration is not performed [26].The challenge remains on having good quality AM radiographic data.In order to enable better comparative materials in the reconciliation phase, clinicians must properly produce and store imaging data, such as radiographs.Interestingly, in 2006 authors had demonstrated that in specific regions of Australia human dentifications were considered as "confirmed" only if AM radiographic records were available.Currently, there is no guideline that guarantees "established", "positive" or even "confirmed" human identifications based on existing radiographs.The problem is that radiographs must not be merely available, but must have good evidentiary value, and this value is strongly associated with the region of interest (ROI) of the radiograph, the spatial resolution of the image, and even how recent the AM radiograph is.In this context, making radiographic reconciliations mandatory is not feasible based on the current evidence and practice.On the other hand, dentists should be encouraged to take and compare radiographs and should be aware of the existing standards for comparative human identification.
In order to avoid technical mistakes on the interpretation of dental radiographs [27], however, forensic odontologists must be aware of image visualization and interpretation.Based on the present findings, it is the duty of this paper to support the inclusion of image interpretation in forensic odontology training programmes.Experts must be aware of the available tools to visualize and navigate through 2D and 3D images.This is to say that the collection of evidence in the form of dental features will be optimized if expert's knowledge is optimal as well.
It must be noted that while, on one hand, therapeutic features enhance the confidence and accuracy of forensic odontologists for human identification, on the other hand preventive dentistry is becoming more popular.Consequently, less therapeutic features might be expected in future human identification procedures.Forensic odontologists showed proper knowledge of the importance of morphological features, but cases with unrestored teeth had higher false negative rates suggesting that radiographic anatomy is an important factor to improve professional performance even more.

Conclusion
Cases with unrestored teeth registered higher rates of false negatives.The combination of therapeutic and morphological (and eventually other) features figure as the best approach to solve more challenging cases.In general, forensic odontologists relied more often on restorations > tooth morphology > and bone features when comparing radiographs for human identification.Additionally, they felt more confident in establishing an identity when dental restorations were present.

Fig. 1 .
Fig. 1.Antemortem (AM) and postmortem (PM) intraoral radiographs used in the simulated casework #1.Legend: The figure illustrates the permanent mandibular right second molar with an occlusal restoration, retained roots of the first molar (present only in the AM radiograph), and missing premolars.The radiographs belong to the same individual and were obtained in different times.

Fig. 2 .
Fig. 2. Antemortem (AM) and postmortem (PM) intraoral radiographs used in the simulated casework #2.Legend: The AM radiograph is a bitewing and shows seven teeththe permanent maxillary left first and second premolars, and second molar, as well as the mandibular first and second premolars and molars.The maxillary first molar is missing.The PM data consists of a periapical radiograph of the permanent maxillary left first and second premolars, and second molar.The maxillary molar is missing.None of the teeth present therapeutic features.The radiographs belong to the same individual and were obtained in different times.

Fig. 3 .
Fig. 3. Antemortem (AM) and postmortem (PM) intraoral radiographs used in the simulated casework #3.Legend: AM and PM radiographs are periapical and show permanent teeth of the right side.Five teeth are observed AM: first and second premolars, and first, second and third molars.Three teeth are observed in PM: first and second premolars and first molar.The radiographs belong to the same individual and were obtained in different times.Explainable discrepancy is observed between the features present on the permanent mandibular right second premolar.

Table 1
Structured questionnaire used to assess experts' opinion in this survey.

Table 2
Experts' responses for AM/PM comparison to each casework-based scenario.

Table 4
Different dental features that were examined in intraoral radiographs and mentioned by the participants.