Human Identification via Lateral Patella Radiographs: A Validation Study,

Evaluating artificial intelligence for comparative radiography

INTRODUCTION

Comparative radiography is a forensic identification and shortlisting technique based on the comparison of skeletal structures in ante-mortem and post-mortem images. The images (e.g., 2D radiographs or 3D computed tomographies) are manually superimposed and visually compared by a forensic practitioner. It requires a significant amount of time per comparison, limiting its utility in large comparison scenarios.

METHODS

We propose and validate a novel framework for automating the shortlisting of candidates using artificial intelligence. It is composed of (1) a segmentation method to delimit skeletal structures' silhouettes in radiographs, (2) a superposition method to generate the best simulated "radiographs" from 3D images according to the segmented radiographs, and (3) a decision-making method for shortlisting all candidates ranked according to a similarity metric.

MATERIAL

The dataset is composed of 180 computed tomographies and 180 radiographs where the frontal sinuses are visible. Frontal sinuses are the skeletal structure analyzed due to their high individualization capability.

RESULTS

Firstly, we validate two deep learning-based techniques for segmenting the frontal sinuses in radiographs, obtaining high-quality results. Secondly, we study the framework's shortlisting capability using both automatic segmentations and superimpositions. The obtained superimpositions, based only on the superimposition metric, allowed us to filter out 40% of the possible candidates in a completely automatic manner. Thirdly, we perform a reliability study by comparing 180 radiographs against 180 computed tomographies using manual segmentations. The results allowed us to filter out 73% of the possible candidates. Furthermore, the results are robust to inter- and intra-expert-related errors.

Radiographic presentation of artifactual dyed hair on lateral cephalograms, chemical processing, and forensic application: Novel case report

Forensic assessment employs an array of methods to identify human remains. Radiologic examinations with panoramic radiographs, computed tomography scans, Waters view, and nuclear magnetic resonance imaging may offer evidentiary clues in challenging cases, such as mass disasters. In these cases, alternative forensic tools are used to narrow lists of target victims using their biological features. This study aims to I) report on the unusual radiographic aspect of chemicals used for hair dyeing, and II) discuss the potential forensic application of this finding for human identification. The case depicts an asymptomatic 14-years-old female who presented for orthodontic therapy. During radiographic examination on a lateral cephalogram, numerous thin radiopaque streaks were visible, extending to the posterior neck between the occipital region of the skull base and vertebra C6. Clinical investigations revealed that these were artifactual hair images (possibly documented for the first time in the scientific literature). Inductively coupled plasma mass spectrometry (ICP-MS) analysis of the patient's scalp hair was performed for 10 heavy metals, including zinc, copper, iron, chromium, nickel, cadmium, tin, lead, antimony, and bismuth. Eight of these metals were detected at normal levels, ranging from 160 parts per million (ppm) for zinc to less than 1 ppm for nickel, cadmium, tin, lead, and antimony. Conversely, slightly elevated levels of chromium at 0.41 ppm and bismuth at 0.025 ppm were found in the hair sample. The distinctive radiographic presentation of artifactual hair images combined with the chemical properties of hair exposed to dye products may provide useful traces for human identification, especially in mass disasters.

Infracranial radiographic comparison for human identification: A study of image quality and tissue shielding effects

In this study, we explore how image resolution and tissue shielding can impact correct classification rates (CCRs) of infracranial radiographic comparisons undertaken using small field-of-view radiographs. Thirty-six identification arrays (using clavicles and seventh cervical vertebra) were constructed with each array comprised of five radiographs: one X-ray of a single dry bone (postmortem [PM] skeletal image) and four simulated antemortem [AM] radiographs (radiographs taken pre-skeletization). One AM radiograph in each array represented the ground truth match to the PM radiograph (=25% rate of randomly selecting the correct match). Radiographs were digitally manipulated, so that four varieties of decreasing blur (Gaussian blur = 12-0 pr across 24 arrays), and, for PM clavicles, four varieties of decreasing hard tissue shielding (opacity of 40-0% across 12 arrays) existed. Arrays were evaluated, for their correct PM/AM pair, by 8 anthropologists trained in chest radiograph comparison (CXR; either currently or formerly competency certified by the Defense POW/MIA Accounting Agency CXR training program), 28 current American Board of Forensic Anthropology (ABFA) diplomates and 30 novices. Analysts' CCRs substantially improved when Gaussian blur was <10 pr (55% CCR at >10 pr vs. 89% for <10 pr). Tissue shielding effects increased CCRs on average by +10% for each -10% opacity step (between 40-0% opacity). The CXR anthropologists were the most tolerant of the more challenging identification contexts (highest blur and opacity), reconfirming that analyst training and expertise is an important factor, especially when poorer quality radiographic images are the subject of analysis.

Infra-cranial radiographic comparison for human identification: A study of analyst expertise

Radiographic comparison for identification is widely utilized. However, these methods are qualitative and subject to analyst ability to correctly read and interpret radiographs. With regards to infra-cranial radiographs, few studies have been conducted to explore the role of practitioner expertize on correct classification rates (CCRs). Here, we undertake two such studies using forensic anthropologists [American Board of Forensic Anthropology (ABFA) certified, practicing but not board-certified anthropologists, and chest radiograph comparison (CXR) anthropologists trained via the Defense POW/MIA Accounting Agency CXR competency training program] and compare their results to novices. To ensure participants only referred to the same skeletal morphology, we cropped radiographs to single bones. An array of four simulated antemortem radiographs was presented to each assessor with each postmortem radiograph. Assessors evaluated arrays for a correct match, which was always present, yielding a 25% rate for random correct selections. Study 1 used anteroposterior C7, posteroanterior second metacarpal, and lateral calcaneus radiographs (three arrays each for nine arrays total), which yielded 86, 81, 69, and 68% mean CCRs for CXR experts, ABFA anthropologists, non-ABFA anthropologists and novices, respectively. Study 2 used anteroposterior C7 and left clavicles (three arrays each for six arrays total), yielding mean CCRs of 100, 96, and 84% for CXR experts, ABFA anthropologists and novices, respectively. As reflected by the CCRs, expertise is clearly a factor for radiographic comparisons, evident not just between novices and anthropologists, but also between anthropologists. We recommend all radiographic comparison analysts be subject to competency/proficiency tests prior to their engagement for forensic casework.

Computerized delineation of the teeth and comparison with a smiling photograph: identification of a body skeletonized by cadaverous ichthyofauna action

Amazonian's forensic experts usually experience considerable difficulties in the identification of drowned bodies rescued from the rivers of the Amazon basin, since they are frequently found skeletonized by the action of the cadaverous ichthyofauna. In these circumstances, especially when the soft tissues are completely absent, bones and teeth may represent the unique source of information for the identification of the body. This work reports a case of positive identification of a body skeletonized by scavenger ichthyofauna. The identification was performed by comparing computerized delineation of the dental characteristics of the cadaver with those observed in a smiling photograph of the victim. This report also highlights the ferocity of Cetopsis candiru (candiru cobra), Cetopsis coecutiens (candiru-açu) and Calophysus macropterus (piracatinga).

Estimation of sex based on patella measurements in a contemporary Chinese population using multidetector computed tomography: An automatic measurement method

Sex estimation is an important part of creating a biological profile, and ultimately assisting in creating a presumptive identification of unidentified skeletal remains. However, manual methods of anthropometric are time-consuming and prone to observer variability. The present study is an attempt to estimation of sex from automatic measurement of patella by multidetector computed tomography (MDCT) in a contemporary Chinese population. Four measurements for every patella, including maximum height (MAXH), maximum breadth (MAXB), maximum thickness (MAXT) and patellar volume (PV), were automatically provided by the software from CT image of 300 Chinese. The sample is composed of 156 males and 144 females with an average age of 41.44 and 45.68 years, respectively. The statistical analyses showed that all variables were sexually dimorphic. Receiver operating characteristic (ROC) analysis was performed to estimate sex from patella. The univariate analysis of each patellar parameter yielded a sex classification accuracy rate of 73.1% to 85.7%. The classification accuracy rates of sex estimation using the combination of the patellar parameters are 81.9% to 91.6%. This paper provides indications that the patella is important bone for sex estimation and it may be used as an alternative in forensic cases when the skull and pelvis are unavailable.

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.

Intra- and Inter-Element Variability in Mitochondrial and Nuclear DNA from Fresh and Environmentally Exposed Skeletal Remains

Successful identification of skeletonized remains often relies upon DNA analyses, frequently focusing on the mid-diaphysis of weight-bearing long bones. This study explored intra-bone DNA variability using bovine and porcine femora, along with calcanei and tali. DNA from fresh and short-term environmentally exposed bone was extracted utilizing demineralization and standard lysis buffer protocols, and DNA quantity and quality were measured. Overall, femoral epiphyses, metaphyses, and the tarsals had more nuclear and mitochondrial DNA than did the femoral diaphyses. DNA loss was much more rapid in buried bones than in surface exposed bones, while DNA quality differed based on environment, but not bone region/element. The demineralization protocol generated more DNA in some bone regions, while the standard lysis was more effective in others, and neither significantly affected DNA quality. Taken together, these findings reinforce the importance of considering inter- and intra-bone heterogeneity when sampling skeletal material for forensic DNA-based identifications.

An Automated Two-dimensional Pairwise form Registration Method for Pair-matching of Fragmented Skeletal Remains

This study introduces an automated pairwise method for osteological pair-matching of fragmented skeletal remains using two-dimensional fragmented outlines extracted from photographs. The form data are used in pairwise iterative closest point registrations with rigid transformations. A modified version of the average Hausdorff distance is calculated to remove any coordinate correspondences with outline fracture margins, which allow the distance analysis of fragmented outlines. A dilation modification to the Hausdorff distance is proposed creating a greater separation between true- and false-pairs. The sample consists of 122 calcanei (61 pairs) from the UI-Stanford collection. Performance statistics are provided for simulated fragmented and complete assemblages. Results indicate up to 98% accuracy for fragmented and complete assemblages. The dilated Hausdorff distance performed similarly across assemblages, but showed a slight decrease in performance for the complete assemblage. This approach provides a useful short listing tool to reduce the number of visual comparisons required in large commingled assemblages.

Elliptical Fourier analysis: fundamentals, applications, and value for forensic anthropology

The numerical description of skeletal morphology enables forensic anthropologists to conduct objective, reproducible, and structured tests, with the added capability of verifying morphoscopic-based analyses. One technique that permits comprehensive quantification of outline shape is elliptical Fourier analysis. This curve fitting technique allows a form's outline to be approximated via the sum of multiple sine and cosine waves, permitting the profile perimeter of an object to be described in a dense (continuous) manner at a user-defined level of precision. A large amount of shape information (the entire perimeter) can thereby be collected in contrast to other methods relying on sparsely located landmarks where information falling in between the landmarks fails to be acquired. First published in 1982, elliptical Fourier analysis employment in forensic anthropology from 2000 onwards reflects a slow uptake despite large computing power that makes its calculations easy to conduct. Without hurdles arising from calculation speed or quantity, the slow uptake may partly reside with the underlying mathematics that on first glance is extensive and potentially intimidating. In this paper, we aim to bridge this gap by pictorially illustrating how elliptical Fourier harmonics work in a simple step-by-step visual fashion to facilitate universal understanding and as geared towards increased use in forensic anthropology. We additionally provide a short review of the method's utility for osteology, a summary of past uses in forensic anthropology, and software options for calculations that largely save the user the trouble of coding customized routines.

Evolution of Forensic Anthropological Methods of Identification

Forensic identification of human remains has long been a core contribution of forensic anthropologists to death investigations. The array and scientific robusticity of the identification methods employed by the anthropologist has evolved in the last several decades, and as with other nonidentification methods, anthropologists have embraced the progression toward the use of validated and statistically defensible methods for identification. This article presents an overview of the role that the forensic anthropologist plays in the identification of human remains and the evolution of anthropological methods of identification.