Validation of post mortem dental CT for disaster victim identification

Forensic odontology in DVI-A path forward

Dental identification is a pillar of disaster victim identification (DVI). Dental identification is accurate, efficient, inexpensive, and accepted in courts of law. The (known) antemortem (AM) dental charts and radiographic images acquired from the dentist of the missing person are evaluated, processed, and compared to post mortem (PM) findings present in the dentition or fragments of the dentition of the deceased individual. These comparisons evaluate and assess individuating restorative dental work, dental anatomical areas of concordance, spatial relationships of teeth one to another, and occasionally calculate the degree of "uniqueness" of either or both of the AM and PM dentition compared to known population databases. In a multiple fatality incident, odontologists may utilize age stratification to assist other means of identification. Computer comparison algorithms using recorded data can indicate possible matches between AM and PM data sets. Following clinical assessment, collection of post mortem tooth specimens for DNA profiling generation may be undertaken. This paper will highlight modern and efficient use of these tools. The framework for how dental identification in these incidents is currently managed is presented. The authors propose a change to this approach that moves away from interpretive subjective assessment toward comparisons based largely on objective data. The aim of this paper is to highlight the benefits of minimizing subjective decisions and maximizing objective data in the dental DVI process while simultaneously reducing risk to clinical personnel and minimizing costs by reducing the number of clinicians required onsite.

Detecting missing teeth on PMCT using statistical shape modeling

The identification of teeth in 3D medical images can be a first step for victim identification from scant remains, for comparison of ante- and postmortem images or for other forensic investigations. We evaluate the performance of a tooth detection approach on mandibles with missing parts or pathologies based on statistical shape models. The proposed approach relies on a shape model that has been built from the full lower jaw, including the mandible and teeth. The model is fitted to the target, resulting in a reconstruction, in addition to a label map that indicates the presence or absence of teeth. We evaluate the accuracy of the proposed solution on a dataset consisting of 76 target mandibles, all extracted from CT images and exhibiting various cases of missing teeth or other cases, such as roots, implants, first dentition, and gap closure. We show an accuracy of approximately 90% on the front teeth (including incisors and canines in our study) that decreases for the molars due to high false-positive rates at the wisdom teeth level. Despite the drop in performance, the proposed approach can be used to obtain an estimate of the tooth count without wisdom teeth, tooth identification, reconstruction of the existing teeth to automate measurements taken as part of routine forensic procedures, or prediction of the missing teeth shape. In comparison to other approaches, our solution relies solely on shape information. This means it can be applied to cases obtained from either medical images or 3D scans because it does not depend on the imaging modality intensities. Another novelty is that the proposed solution avoids heuristics for the separation of teeth or for fitting individual tooth models. The solution is therefore not target-specific and can be directly applied to detect missing parts in other target organs using a shape model of the new target.

Validity of postmortem computed tomography for use in forensic odontology identification casework

Forensic Odontology (FO) identification compares antemortem (AM) and postmortem (PM) dental datasets and is widely accepted as a primary identifier. Traditionally, a PM dental examination is undertaken in the same manner as a dental examination conducted for a living patient. Recently, the increased forensic application of computed tomography (CT) offers an alternative source of PM data. While charting from PMCT is widely accepted as less accurate, the impact on reconciliation is unknown. This study aims to determine if reconciliation outcome differs when PM dental data is collected from PMCT, compared with conventional PM examination. PMCT data was reviewed for 21 cases previously completed using conventional PM dental examination. Operators blinded to original identification outcomes charted from CT images before comparing to AM data to form an opinion regarding identity. Opinions formed were compared with original identification outcomes. Differences in PM dental charting between the two methods and the evidentiary value of AM and PM datasets were assessed to determine driving factors of differences in identification outcome. Compared to conventional PM dental examination, PMCT examination resulted in similar or less certain identification outcomes. Discrepancies in outcome were driven by the quality of AM and PM datasets rather than inaccuracies in charting from PMCT. Based on the results of this study, both conventional and PMCT methods of PM dental examination can reach similar identification outcomes. However, operators remained more certain in establishing identity when conducting conventional PM dental examinations especially when AM data was lacking.

Transformer based tooth classification from cone-beam computed tomography for dental charting

Dental charting is a useful tool in physical examination, dental surgery, and forensic identification. However, manual dental charting faces some difficulties, such as inaccuracy and psychiatric burden in forensic identification. As a critical step of dental charting, tooth classification can be completed on dental cone-beam computed tomography (CBCT) automatically to solve the above difficulties. In this paper, we build a deep neuron network which accepts a 3D CBCT image patch that contains the region of interest (ROI) of a tooth as input and then outputs the type of the tooth. Although Transformer-based neural networks outperform CNN-based neural networks in many natural image processing tasks, they are difficult to apply to 3D medical images. Therefore, we combine the advantages of CNN and Transformer structure to improve the existing methods and propose the Grouped Bottleneck Transformer to overcome the drawbacks of the Transformer, namely the requirement of large training dataset and high computational complexity. We conducted an experiment on a clinical data set containing 450 training samples and 104 testing samples. Experiments show that our network can achieve a classification accuracy of 91.3% and an AUC score of 99.7%. To further evaluate the effectiveness of our method, we tested our network on the publicly available medical image classification dataset MedMNIST3D. The result shows that our network outperforms other networks on 5 out of 6 3-dimensional medical image subsets.

A pilot study of an automated personal identification process: Applying machine learning to panoramic radiographs

Purpose

This study aimed to assess the usefulness of machine learning and automation techniques to match pairs of panoramic radiographs for personal identification.

Materials and Methods

Two hundred panoramic radiographs from 100 patients (50 males and 50 females) were randomly selected from a private radiological service database. Initially, 14 linear and angular measurements of the radiographs were made by an expert. Eight ratio indices derived from the original measurements were applied to a statistical algorithm to match radiographs from the same patients, simulating a semi-automated personal identification process. Subsequently, measurements were automatically generated using a deep neural network for image recognition, simulating a fully automated personal identification process.

Results

Approximately 85% of the radiographs were correctly matched by the automated personal identification process. In a limited number of cases, the image recognition algorithm identified 2 potential matches for the same individual. No statistically significant differences were found between measurements performed by the expert on panoramic radiographs from the same patients.

Conclusion

Personal identification might be performed with the aid of image recognition algorithms and machine learning techniques. This approach will likely facilitate the complex task of personal identification by performing an initial screening of radiographs and matching ante-mortem and post-mortem images from the same individuals.

Prevalence of calcified epiglottis in postmortem computed tomography. Is there a correlation to failed endotracheal intubation?

OBJECTIVES

Calcification of the epiglottis is a normal physiological degenerative process, although it can also be a consequence of infection or trauma. There are three possible forensically relevant consequences from epiglottic calcification: misinterpretation as foreign bodies, dysphagia as a major contributing factor to aspiration, and association with difficult intubation or a misplaced ventilation tube. It is the aim of this study (I) to inquire about the prevalence of epiglottic calcification in postmortem CT in general and (II) to investigate whether calcification of the epiglottis is linked to a higher incidence of failed endotracheal intubation.

METHODS

We retrospectively analysed 2930 consecutive cases in postmortem CT at the Institute of Forensic Medicine.

RESULTS

The prevalence of epiglottic calcification was 4.1%. Higher age and male sex are associated with an increased risk of epiglottic calcification. There was no calcification of the epiglottis in the cases with misplacement of the ventilation tube in the esophagus.

CONCLUSIONS

To verify the result of our study, that is, the calcification of the epiglottis is not linked to a higher incidence of failed endotracheal intubation, it might be reasonable to repeat this study with a more representative study population. The high interindividual variations of calcified epiglottis could be used for identification.

Forensic odontology identification response to terrorist attacks in Paris November 2015

The terrorist attacks of November 2015 led to the immediate death of 129 victims admitted to the Legal and Forensic Medicine Institute of Paris, including 41 unidentified. During the Disaster Victim Identification (DVI) operations, 22 bodies were examined by the postmortem (PM) dental team with the aim of establishing PM odontograms. At the same time, the dental expert in the antemortem (AM) unit collected a large number of dental files, progressively filtered as the list of missing persons became reduced. Feedback from these events has highlighted the difficulties of implementing the DVI chain principles in a legal framework, published the day before the attacks, and also the technical complexity of collecting dental data on a week end of terror. The return on experience after this event has represented a paradigm shift on previous methods of DVI in Paris and even more in France. Indeed, the victim identification procedure was redesigned, integrating new technical means such as a CT scan directly on spot, allowing the extraction of maxillofacial data as soon as possible in order to support the PM dental examination team. Moreover, the National Dental Council proceeded to the overall remodeling of the dental identification unit, which is composed of trained members, from local, regional and national aspects. These forensic experts are dedicated, at the request of the legal authorities, to DVI operations and deployed throughout the country capable of managing AM and PM data. This unit aims also to share experiences and awareness-raising among health professionals and investigators in order to optimize a better submission of AM elements and also to enhance the major interest of odontology as a primary identifier in disaster.

Mastoid, middle ear and inner ear analysis in CT scan - a possible contribution for the identification of remains

Objective

We aimed to evaluate whether the internal structures of the human ear have anatomical characteristics that are sufficiently distinctive to contribute to human identification and use in a forensic context.

Materials and methods

After data anonymisation, a dataset containing temporal bone CT scans of 100 subjects was processed by a radiologist who was not involved in the study. Four reference images were selected for each subject. Of the original sample, 10 examinations were used for visual comparison, case by case, against the dataset of 100 patients. This visual assessment was performed independently by four observers, who evaluated the anatomical agreement using a Likert scale (1–5). Inter-observer agreement, true positive rate, positive predictive value, true negative rate, negative predictive value, false positive rate, false negative rate and positive likelihood ratio (LR+) were evaluated.

Results

Inter-observer agreement obtained an overall Cohen’s Kappa = 99.59%. True positive rate, positive predictive value, true negative rate and negative predictive value were all 100%.

Conclusion

Visual assessment of the mastoid examinations was shown to be a robust and reliable approach to identify unique osseous features and contribute to human identification. The statistical analysis indicates that regardless of the examiner’s background and training, the approach has a high degree of accuracy.

Odontological identification dental charts based upon postmortem computed tomography compared to dental charts based upon postmortem clinical examinations

Since the beginning of this century, the use of postmortem computed tomography (PMCT) in forensic autopsies has increased. In this study we examined how similar dental charts created using PMCT as a solitary examination mode were to dental charts created using the conventional method of a clinical inspection including intraoral radiographs. A total of 100 previously performed dental identification cases were retrospectively included in the study. For each case, a dental chart was created solely based upon PMCT. The PMCT based dental chart was subsequently compared with the chart created from the previous conventional identification examination. Based upon the accuracy, sensitivity and specificity values PMCT performed very well compared to the conventional method in the identification concerning presence or absence of teeth, the presence of crowns, bridges and endodontic treatments as well as the presence and types of fillings. PMCT performed poorly concerning the extension of fillings and identification of small, tooth-colored fillings. The use of PMCT is a valuable supplement to the conventional methods available for forensic odontologists and may be of great value for initial screening in mass fatalities.

Forensic odontology in DVI: current practice and recent advances

Forensic odontology frequently plays a significant role in identification of the victims of multi-fatality disasters, but not in all. It depends on adequate dental remains surviving the disaster and on the availability of dental records to be successful. This paper describes current practice in the techniques of identification in forensic odontology and outlines recent advances that are moving into the mainstream.Key PointsForensic odontology plays a key role in mass disaster victim identification (DVI) when good-quality antemortem (AM) dental records are available.Images including radiographs, computerized tomography (CT) data and three-dimensional (3D) scan data are considered more reliable AM records than written dental charts and odontograms.Interpretation, transcription and comparison of dental datasets are complex processes that should be undertaken only by trained dental professionals.The future of forensic odontology DVI techniques is likely to include the use of 3D datasets for comparison.

Post-mortem computed tomography as part of dental identification - a proposed guideline

PURPOSE

This paper presents a proposed guideline for the use of post-mortem computed tomography (PMCT) during forensic dental identification. Currently, whole-body PMCT is widely used prior to autopsies for the diagnosis of fractures, organ changes, hemorrhages, and for the localization of foreign bodies, but it may also facilitate the odontological identification process in single cases and in cases involving multiple fatalities. Several studies have described the use of PMCT in forensic odontological work, but we have not found any comprehensive set of guidelines on how to perform a forensic odontological examination using PMCT. The aim was to develop guidelines for creating post-mortem dental charts during forensic odontological identification examinations using the standard functions of PMCT.

METHODS

A proposed guideline was developed from 15 selected cases examined at the Section of Forensic Pathology, Department of Forensic Medicine at the University of Copenhagen in Denmark from October 2011 to May 2012. Using the functionalities and three-dimensional (3D) reconstructions of OsiriX DICOM-viewer software (Pixmeo Sarl, Bernex, Geneva, Switzerland) we adjusted the contrast and brightness settings and developed a proposed guideline for creating PMCT-based dental charts. A four-step guideline was produced.

CONCLUSION

In our casework, we are currently using the guidelines proposed herein. The use of PMCT has allowed us to target our clinical examinations, greatly improving their efficiency. Furthermore, PMCT allows the storage of data for later documentation and research. Further research is needed to validate the proposed guideline.

Post-mortem X-ray computed tomography (PMCT) identification using ante-mortem CT-scan of the sphenoid sinus

PURPOSE

To evaluate forensic identification of individuals through visual comparison of sphenoid sinus anatomical configuration using ante- and post-mortem CT-scans.

METHOD AND MATERIALS

Ante- and post-mortem head CT-scan of 33 individuals were retrospectively collected. Ten head CT-scans were randomly selected from various neurological contexts and added to the ante-mortem group. Ten other head CT-scans were randomly selected from our post-mortem PACS and added to the post-mortem group. These CT-scans were assigned into 2 groups for analysis: an ante-mortem group (33 + 10) and a post-mortem group (33 + 10). For ethics and to avoid identification bias, CT-scans were anonymized - not showing any head structure but only sphenoid sinuses. An anatomical based classification system using the sphenoid sinuses anatomical variations was created according to anatomical and surgical literature. This classification was used by readers to identify in two different steps a maximum of matched and then unmatched scans.

RESULTS

The first reader had a sensitivity of 100% [CI: 89.4%-100%] and a specificity of 100% [CI: 99.8%-100%]. Sensitivity and specificity were respectively 93.9% [CI: 79.8%-99.3%] and 99.9% [CI: 99.6%-100%] for the second reader. Positive and negative predictive values were respectively 100% [CI: 89.4%-100%] and 100% [CI: 99.8%-100%] for the first reader. Positive and negative values were respectively 96.9% [CI: 83.8%-99.9%] and 99.9% [CI: 99.7%-100%] for the second reader. Inter-reader variability was estimated by Cohen's kappa and an excellent agreement was found.

CONCLUSION

We reported an excellent validity and reliability of subjective visual comparison of ante- and post-mortem CT-data using an anatomical based classification of the sphenoid sinus.

Classification of teeth in cone-beam CT using deep convolutional neural network

Dental records play an important role in forensic identification. To this end, postmortem dental findings and teeth conditions are recorded in a dental chart and compared with those of antemortem records. However, most dentists are inexperienced at recording the dental chart for corpses, and it is a physically and mentally laborious task, especially in large scale disasters. Our goal is to automate the dental filing process by using dental x-ray images. In this study, we investigated the application of a deep convolutional neural network (DCNN) for classifying tooth types on dental cone-beam computed tomography (CT) images. Regions of interest (ROIs) including single teeth were extracted from CT slices. Fifty two CT volumes were randomly divided into 42 training and 10 test cases, and the ROIs obtained from the training cases were used for training the DCNN. For examining the sampling effect, random sampling was performed 3 times, and training and testing were repeated. We used the AlexNet network architecture provided in the Caffe framework, which consists of 5 convolution layers, 3 pooling layers, and 2 full connection layers. For reducing the overtraining effect, we augmented the data by image rotation and intensity transformation. The test ROIs were classified into 7 tooth types by the trained network. The average classification accuracy using the augmented training data by image rotation and intensity transformation was 88.8%. Compared with the result without data augmentation, data augmentation resulted in an approximately 5% improvement in classification accuracy. This indicates that the further improvement can be expected by expanding the CT dataset. Unlike the conventional methods, the proposed method is advantageous in obtaining high classification accuracy without the need for precise tooth segmentation. The proposed tooth classification method can be useful in automatic filing of dental charts for forensic identification.