The reliability of digitized radiographs for dental identification: a Web-based study

In the era of Daubert and other judicial rulings pertaining to the acceptability of forensic evidence, it is increasingly important that experts are able to testify that their methods have been scientifically tested and that error rates and other factors relating to reliability have been published. The purpose of this study was to determine the reliability of digitized radiographic comparisons for the purposes of dental identification. Participants with various forensic backgrounds and experience levels were passively recruited to the website. Ten forensic identification cases composed of antemortem and postmortem dental radiographs were supplied to examiners using a bespoke website. Participants responded to the cases on two occasions after a one-month washout interval using the ABFO conclusion levels for forensic identifications. A total of 115 first attempts and 87 matched second attempts were received. Of the total responses, 72% were dentally trained respondents who had completed at least one forensic identification case; of these, 38% were experienced forensic dentists who had completed more than 25 identifications. Data relating to accuracy, intra- and inter-examiner agreement, and the effect of case difficulty are presented. Mean accuracy was 85.5% for all cases, with the experienced forensic dentists obtaining a 91% success rate. The inter-examiner agreement on the negative identification cases was classified as poor. The data suggest that dental identifications resulting from the comparison of postmortem and antemortem radiographs are valid, accurate, and reliable when undertaken by experienced odontologists.     

Antemortem dental records: attitudes and practices of forensic dentists

This study was designed to provide insight concerning the attitudes and practices of forensic dentists regarding antemortem dental records reviewed for purposes of dental identification. Forensic dentists were invited to participate in a 10 item survey. The majority of the respondents reported a considerable amount of experience in dental identifications of the deceased. Sixty-six percent reported having suspected dental negligence or fraud in their antemortem record reviews. Only 17% believe that a forensic dentist should report suspected dental negligence, while 31% agree that dental fraud should be reported. Their response to additional issues addressed in the study suggests diversity in the practices and attitudes of forensic dentists in the use of antemortem dental records. In conclusion, opening a dialogue among practicing forensic dentists may lead to a standardized set of recommendations by the appropriate societies in the forensic dental community.     

Associated postmortem dental findings as an aid to personal identification.

The use of the unique features of the human dentition to aid in personal identification is well accepted within the forensic field. Indeed, despite advances in DNA and other identification methodologies, comparative dental identifications still play a major role in identifying the victims of violence, disaster or other misfortune. The classic comparative dental identification employs the use of postmortem and antemortem dental records (principally written notes and radiographs) to determine similarities and exclude discrepancies. In many cases the tentative identification of the individual is unknown and therefore antemortem records cannot be located. In such a situation a dental profile of the individual is developed to aid the search for the individual's identity. With such a profile a forensic odontologist can identify and report indicators for age at time of death, race (within the four major ethnic groups) and sex. In addition to these parameters the forensic dentist may be able to give more insight into the individual. This paper outlines, for the non-expert, some of the additional personal information that can be derived from the teeth of the deceased, and which may assist in their ultimate identification.     

Self-assessment of the forensic value of dental records.

One of the most important aspects of a person's dental record may well prove to be it's potential value should the forensic dental identification of their remains become necessary. The better the quality of the antemortem dental records, the easier and faster the identification of the remains will be. The forensic dentist must be able to select identifying features by decoding the deceased's antemortem dental records. A study was conducted on two groups of dentists who were asked to self-assess the forensic dental value of the dental records maintained in their own practices. The three most frequently recorded identifying dental features, other than caries and restorations, were the presence of diastemas, displaced or rotated teeth, and dental anomalies. Surveyed dentists imbedded identifying information into the removable prosthetic devices fabricated for their patients an average of only 64% of the time. Only 56% of the two groups combined felt that their dental chartings and written records would be extremely useful in dental identifications. It is concluded that the quality of antemortem dental records available for comparison to postmortem remains varies from inadequate to extremely useful. Practicing dentists can become valuable members of the dental identification process by developing and maintaining standards of record keeping which would be valuable in restoring their patients' identity.     

Radiographic human identification using bones of the hand: a validation study

The 1993 Supreme Court case Daubert v. Merrell-Dow Pharmaceuticals, Inc. underscores the importance of validating forensic science techniques. This research examines the validity of using posterior-anterior radiographs of the hand to make positive identifications of unknown human remains. Furthermore, this study was constructed to satisfy the requirements of Daubert's guidelines of scientific validity by establishing a standard methodology for hand radiograph analysis, testing the technique, and noting rates of error. This validation study required twelve participant examiners from the forensic science community, working independently, to attempt to match 10 simulated postmortem radiographs of skeletonized hands to 40 simulated antemortem radiographs of fleshed cadaver hands. The overall accuracy rate of the twelve examiners was 95%, while their collective sensitivity and specificity were 95% and 92%, respectively. However, the accuracy of each examiner was related to the amount of radiological training and experience of the observer. Six Ph.D. forensic anthropologists and four experienced forensic anthropology graduate students correctly identified all the matches. Participant examiners noted bone morphology, trabecular patterns of the proximal and middle phalanges, and distinctive radiopaque and radiolucent features as the anatomical features that aided the identification process. The hand can be an important skeletal element for radiographic positive identification because it contains 27 individual bones for comparative analysis.     

Validation of chest X-ray comparisons for unknown decedent identification

Comparing skeletal structures between antemortem and postmortem chest radiographs is widely used by forensic specialists from many disciplines to positively identify unknown decedents. However, validity assessments of this method have been fairly limited. This study had three objectives: 1) to quantify the reliability of ante- and postmortem chest radiograph comparison for decedent identification; 2) to identify useful radiologic features supporting decedent identification; and 3) to recognize sources of error in decedent identification related to use of comparative radiographs. A forensic pathologist, a forensic anthropologist, and two radiologists participated in the study. Our results showed that chest radiograph comparisons proved reliable, if basic decedent information was provided, and antemortem and postmortem radiographs were adequately positioned and exposed. A "morphological approach" using normal anatomical structures for comparison may provide the most efficient method for accurate identification.     

Skeletal identification using the frontal sinus region: a retrospective study of 39 cases

The importance of identification using the frontal sinus has been previously demonstrated in case reports. In this study, 39 cases of identification using frontal sinus comparison from the Ontario Chief Coroner's Office were reviewed and differences between antemortem and postmortem radiographs examined. All cases involved decedents older than twenty years. Three cases were rejected due to poor antemortem and postmortem film quality. One subject had no frontal sinus. Thirty-five cases provided conclusive postmortem to antemortem pattern matches. Sixteen cases also yielded metric (quantitative) matches. Duration between antemortem and postmortem radiographic examinations, age, gender, and cause of death did not affect the ability to obtain a match. This is the largest study undertaken on actual cases and demonstrates the validity of frontal sinus pattern matching for forensic identification.     

Root Morphology and Anatomical Patterns in Forensic Dental Identification: A Comparison of Computer‐Aided Identification with Traditional Forensic Dental Identification*

Abstract: An online forensic dental identification exercise was conducted involving 24 antemortem–postmortem (AM–PM) dental radiograph pairs from actual forensic identification cases. Images had been digitally cropped to remove coronal tooth structure and dental restorations. Volunteer forensic odontologists were passively recruited to compare the AM–PM dental radiographs online and conclude identification status using the guidelines for identification from the American Board of Forensic Odontology. The mean accuracy rate for identification was 86.0% (standard deviation 9.2%). The same radiograph pairs were compared using a digital imaging software algorithm, which generated a normalized coefficient of similarity for each pair. Twenty of the radiograph pairs generated a mean accuracy of 85.0%. Four of the pairs could not be used to generate a coefficient of similarity. Receiver operator curve and area under the curve statistical analysis confirmed good discrimination abilities of both methods (online exercise = 0.978; UT‐ID index = 0.923) and Spearman’s rank correlation coefficient analysis (0.683) indicated good correlation between the results of both methods. Computer‐aided dental identification allows for an objective comparison of AM–PM radiographs and can be a useful tool to support a forensic dental identification conclusion.     

Reanalysis of Korean War Anthropological Records to Support the Resolution of Cold Cases

Re‐investigation of previously unidentified remains from the Korean War has yielded 55 new identifications, each with corresponding records of prior anthropological analyses. This study compares biological assessments for age at death, stature, and ancestry across (i) anthropological analyses from the 1950s, (ii) recent anthropological analyses of those same sets of remains, and (iii) the reported antemortem biological information for the identified individual. A comparison of long bone measurements from both the 1950s and during reanalysis is also presented. These comparisons demonstrate commonalities and continuing patterns of errors that are useful in refining both research on Korean War cold case records and forensic anthropological analyses performed using methods developed from the 1950s identifications.     

Identification problems in emergencies with human victims

The problems and errors in personal identifications in rescue work after accidents with a large number of human victims are considered. The order of actions of forensic medical experts in personal identification in emergency situation is proposed.     

Dental identification by comparison of antemortem and postmortem dental radiographs: influence of operator qualifications and cognitive bias

Dental forensic identifications based on comparison of antemortem and postmortem radiographs provide effective and reliable evidence. There are no standardized procedures for assessing similarities between different types of dental radiographs (e.g. orthopantomograms, bitewings, and periapical radiographs), and the operator's subjective judgment can considerably affect identification. The purpose of this study is to evaluate the potential influence of experts' qualifications, training, and cognitive bias on the accuracy of identification. Seventy-eight differently qualified and experienced experts underwent an identification test. The expert sample was composed of 10 specialists in emergency care (ER), 10 specialists in legal medicine (ML), 20 pregraduate dental students (STU), 12 dentists (DENT), 20 dentists educated in forensic odontology (DENT-TRA), and 6 experienced forensic odontologists (FOR). The simulated cases required participants to assess the possible matching of 42 postmortem intraoral radiographs with 16 antemortem panoramic radiographs. Accuracy and specificity for the different operator groups were as follows: ER, 0.76-0.70; ML, 0.76-0.88; STU, 0.89-0.82; DENT, 0.87-0.97; DENT-TRA, 0.88-0.92; and FOR, 0.97-1. As evidenced by high rates of accuracy and repeatability, the most experienced forensic odontologist consistently outperformed operators less or differently educated and trained, especially for difficult cases. In our sample, the dentists who received additional education in forensic odontology did not necessarily perform better than dentists who had not received this additional education. Some cognitive bias, mainly the so-called observer effect, emerged as a possible source of outcome variability among the operator groups.     

Identification of incinerated root canal filling materials after exposure to high heat incineration

With the increase in global terrorism there is a higher probability of having to identify victims of incineration events secondary to incendiary explosive devices. The victims of incineration events challenge forensic odontologists when coronal restorations are no longer present to compile postmortem data. With 40 million root canals being completed annually in the United States, a very large pool of antemortem data is available to the forensic odontologist to make positive identifications. When complete and thorough dental records exist, individuals that have undergone surgical and nonsurgical root canal therapy may have materials present in the canal that may aid in identification. This study provides elemental fingerprints of root canal obturation materials to be utilized as a forensic identification aid. This study used scanning electron microscopy/energy dispersive X-ray spectroscopy (SEM/EDS) to assess the elemental composition of materials before and after high temperature incineration. Sixteen endodontic materials were analyzed pre-incineration and placed in extracted teeth. The filled teeth were subjected to incineration at 900 degrees C for 30 min to simulate incineration events or cremation. Incinerated materials were radiographed and re-analyzed to determine if they retained their original elemental composition. Endodontic sealers, gutta percha, root-end filling materials, silver points, and separated files were distinguishable in the canal and traceable after incineration. The authors present a fingerprint of the endodontic obturation materials that are capable of withstanding high heat incineration to be used as an aid for postmortem identification. This work represents the initial stage of database generation for root canal filling materials for use as an aid in forensic identification.     

Anthropological and Radiographic Comparison of Antemortem Surgical Records for Identification of Skeletal Remains*

Abstract: This case review illustrates the important contributions of forensic archeological methods and forensic anthropological analysis to the identification of found skeletal remains. After reassociation of skeletal remains found in two locations, anthropological analysis provided the basis for a presumptive identification and a request for antemortem medical records. Partial DNA profiles were supportive but not conclusive and antemortem dental records were not available. Comparison of antemortem traumas, skeletal morphology, and surgical artifacts with antemortem radiographs and surgical records led to positive identification of an individual missing for almost a decade.     

A Radiographic Study on the Utility of Cranial Vault Outlines for Positive Identifications

A standard method for positive identification is the use of antemortem and postmortem radiographic comparisons. The purpose of this research is to test the visual accuracy of antemortem and postmortem radiographic comparisons of cranial vault outlines and to evaluate their uniqueness using geometric morphometric methods. A sample of 106 individuals with varying levels of education and forensic case experience participated in a visual accuracy test. Of the 106 individuals, only 42% correctly assigned all of the radiographs, with accuracy rates ranging from 70 to 93% for each radiographic comparison. Vault shape was further examined using elliptic Fourier analysis, and paired t‐tests were computed on the first 10 principal components accounting for 100% of the variance, which found no significant differences. The visual accuracy test and elliptic Fourier analysis shows that vault outlines may not be unique enough for positive identifications when used as a sole indicator.     

Skeletal identification by radiographic comparison: blind tests of a morphoscopic method using antemortem chest radiographs

This study investigated the value of antemortem (AM) and postmortem (PM) radiographs of the claviculae and C3-T4 vertebrae to identify skeletons of missing U.S. soldiers from past military operations. In total, 12 field-recovered skeletons and AM chest radiographs of 1460 individuals were used. For each skeleton, examiners analyzed an array of AM chest radiographs (up to 1000 individuals) and attempted to identify the correct PM/AM radiographic match. When examiners were able to compare all images within a single test, only true-positive identifications were made. When AM radiographs were presented one-at-a-time, in sequential order, and without examiners having knowledge of array size, erroneous identifications resulted but they were almost exclusively made by untrained examiners (accuracy = 35% vs. 90% for trained examiners). This study demonstrates the value of chest radiographs for the identification of disarticulated and even eroded skeletons, but only when methods are wielded by trained examiners.     

The Reliability of Facial Recognition of Deceased Persons on Photographs

In humanitarian emergencies, such as the current deceased migrants in the Mediterranean, antemortem documentation needed for identification may be limited. The use of visual identification has been previously reported in cases of mass disasters such as Thai tsunami. This pilot study explores the ability of observers to match unfamiliar faces of living and dead persons and whether facial morphology can be used for identification. A questionnaire was given to 41 students and five professionals in the field of forensic identification with the task to choose whether a facial photograph corresponds to one of the five photographs in a lineup and to identify the most useful features used for recognition. Although the overall recognition score did not significantly differ between professionals and students, the median scores of 78.1% and 80.0%, respectively, were too low to consider this method as a reliable identification method and thus needs to be supported by other means.     

Identification of the Edentulous Individual: An Investigation into the Accuracy of Radiographic Identifications*

Abstract: The dental identification of edentulous individuals can be challenging based on the lack of antemortem materials and unique features visible on radiographs. The constant resorption of the alveolar ridges further complicates the process. The purpose of this study was to quantify the error rate and reliability of dental identifications based on a comparison of synthesized antemortem and postmortem radiographs of edentulous individuals. Ten observers examined ten cases on two occasions and reported dichotomous and conclusion level decisions. These were analyzed using Kappa and Receiver Operator Characteristics. The mean area under the curve was 0.75 and the mean sensitivity was 0.57 and specificity was 0.83. These results suggest that dental identifications of edentulous individuals using radiographs alone have a high error rate and should be dual reported. These data add further weight to the argument that all dental prostheses should be labeled.     

A new algorithm for use in computer identification

On 9 May 1987, a Soviet-made IL-62M Polish airliner, LOT Flight 5055, crashed, exploded, and burned, killing the crew and 183 passengers. A forensic science team from the Armed Forces Institute of Pathology, comprised of 6 dental officers, 3 forensic pathologists, and 3 medical photographers, worked in concert with the Polish forensic science team. The small number of antemortem records and the extreme fragmentation of the remains presented a new scenario for computer use. Typically, the Computer-Assisted Postmortem Identification (CAPMI) software is used to compare remains against an antemortem database. Results are listed by the number of tooth-to-tooth matches based on restorative or other characteristics or both. The Polish disaster confounded this approach to some degree, however, and suggested a reconsideration of the theory on which the sort is made, that is, that the cases with maximum number of matches to preexisting dental records would be the most likely identification (ID) match. A hypothesis was constructed that, if searches were accomplished for fragments with a minimum number of mismatches, the correct matches would appear higher in the rank order. Six antemortem records (that had all dental information) were sorted against one hundred and twelve postmortem fragmented records. The resulting report was reordered so that records were listed by minimum number of mismatches. There was significant improvement in rank placement for all of the records. Thus it was accepted that in the situation of highly fragmented remains a different sorting based on the number of mismatches is indicated. Programming changes to make this option available have been implemented in the new version of CAPMI.     

Interpretation,confidence and application of the standardised terms: Identified,Probable, Possible,Exclude and Insufficient in forensic odontology identification

Forensic odontology identification scales are used to express certainty of identifications of deceased persons. These standardized scales are assumed to convey unambiguous expert opinions and facilitate communication between forensic odontologists and end users. However, to date no studies have investigated how the experts interpret and use these scales.Forensic odontology identification scales are used to express certainty of identifications of deceased persons. These standardized scales are assumed to convey unambiguous expert opinions and facilitate communication between forensic odontologists and end users. However, to date no studies have investigated how the experts interpret and use these scales.This paper aims to examine the interpretation of the DVISYS forensic identification scale and choices of the levels in the scale subsequent to, and derived from, comparison of pairs of dental radiographs by extending the analysis of the data collected in the study by Page and Lain et. al. 2017.The studied variables: self-reported confidence, forced binary decision of match and non-match, choice of level in the DVISYS scale (Identified, Probable, Possible, Insufficient and Exclude) were further analysed in this study using mixed models for relationships between the choices of level in the identification scale and the fundamental beliefs of likelihood of identification.The results of this further analysis showed that the reported confidence of the decisions was correlated to the difficulty of cases, and as confidence decreased the use of less definitive terms (‘Probable’, ‘Possible’ and ‘Insufficient’) increased. ‘Probable’ and ‘Possible’ were used mainly in underlying beliefs below that of ‘Identified’ whereas ‘Insufficient’ was used mainly to convey a sublevel of ‘Exclude’. The use of ‘Insufficient’ in this study was not consistent with the prescribed definition of the term.The participants of the original study were not aware of the difficulty grading of the cases nor were required to grade them, however the reported confidence was systematically correlated to difficulty. Furthermore, indicated confidence level was correlated with choice of level on the scale in general, but the interpretation of the definition and application of the terms varied.The findings reported here contribute to the foundational knowledge of factors governing the interpretation and application of the DVISYS forensic odontology identification scale and suggest that this scale may need to be modified.     

Frequencies of morphological characteristics in two contemporary forensic collections: implications for identification

Positive identification relies on comparison of antemortem and postmortem data. Some identifications are based on morphological features such as fracture, pathological condition, and surgical hardware, despite little literature indicating the frequencies of such traits. This study examines whether such features are sufficiently rare as to be deemed individualizing. Data were collected on two modern North American skeletal collections (N=482 individuals). Presence/absence of features was scored by skeletal element and side. Results indicate that frequencies vary by geographic region (higher frequency of fractures and pathological conditions in New Mexico while individuals in Tennessee were more likely to have surgical interventions), many features such as fractures are remarkably common and that even suites of traits may not be individualizing. Caution is warranted when using written data rather than radiographic comparisons as the primary source of identification. The implications of these findings to missing person databases are also discussed.