Accuracy of dental identification of individuals with unrestored permanent teeth by visual comparison with radiographs of mixed dentition

Forensic dentistry plays a major role in human identification. Teeth carry individual characteristics that differ among different individuals. Dental radiographs depict reality objectively, being the most reliable tool for dental identification. The first aim of this study was to evaluate the accuracy of dental identification of individuals with permanent unrestored teeth by visual comparison with radiographs of mixed dentition. The second aim was to learn which anatomical features were compared by examiners with different backgrounds. A total of 19 forensic experts participated in a web-based questionnaire to assess identification of 12 simulated cases; each case required the radiographic comparison of 1 dental PM radiograph to 3 dental AM radiographs, of which only one was the correct match. The examiners were given four options following the ABFO guidelines: established identification, possible identification, insufficient data and exclusion; the participants also explained the reason for each of their conclusions. The accuracy of the methodology was 75,4%, the sensitivity was 53,5% and the specificity was 86,4%. Overall, there was a tendency of the observers to overlook non-dental characteristics. Not surprisingly, dental identification by visual comparison of radiographs was not immune to subjectivity and, even analysing the same category of features, different conclusions and consequently different percentages of accuracy were reached. When matching the correct AM radiograph, most examiners compared the root morphology of the first molars and the shape of the maxillary sinus. When one of the AM radiographs was not matched, the examiners mostly asserted that there was insufficient data to reach a conclusion due to the lack of distinctive and comparable features. With AM and PM radiographs showing different development stages, accuracy was correlated to the age of the AM radiograph. (c) 2018 Elsevier B.V. All rights reserved.