A Test of the DSP Sexing Method on CT Images from a Modern French Sample

Automatic variable extraction from 3D coxal bone models for sex estimation using the DSP2 method

Thanks to technical progress and the availability of virtual data, sex estimation methods as part of a biological profile are undergoing an inevitable evolution. Further reductions in subjectivity, but potentially also in measurement errors, can be brought by approaches that automate the extraction of variables. Such automatization also significantly accelerates and facilitates the specialist's work. The aim of this study is (1) to apply a previously proposed algorithm (Kuchař et al. 2021) to automatically extract 10 variables used for the DSP2 sex estimation method, and (2) to test the robustness of the new automatic approach in a current heterogeneous population. For the first aim, we used a sample of 240 3D scans of pelvic bones from the same individuals, which were measured manually for the DSP database. For the second aim a sample of 108 pelvic bones from the New Mexico Decedent Image Database was used. The results showed high agreement between automatic and manual measurements with rTEM below 5% for all dimensions except two. The accuracy of final sex estimates based on all 10 variables was excellent (error rate 0.3%). However, we observed a higher number of undetermined individuals in the Portuguese sample (25% of males) and the New Mexican sample (36.5% of females). In conclusion, the procedure for automatic dimension extraction was successfully applied both to a different type of data and to a heterogeneous population.

Sex estimation from coxal bones using deep learning in a population balanced by sex and age

In the field of forensic anthropology, researchers aim to identify anonymous human remains and determine the cause and circumstances of death from skeletonized human remains. Sex determination is a fundamental step of this procedure because it influences the estimation of other traits, such as age and stature. Pelvic bones are especially dimorphic, and are thus the most useful bones for sex identification. Sex estimation methods are usually based on morphologic traits, measurements, or landmarks on the bones. However, these methods are time-consuming and can be subject to inter- or intra-observer bias. Sex determination can be done using dry bones or CT scans. Recently, artificial neural networks (ANN) have attracted attention in forensic anthropology. Here we tested a fully automated and data-driven machine learning method for sex estimation using CT-scan reconstructions of coxal bones. We studied 580 CT scans of living individuals. Sex was predicted by two networks trained on an independent sample: a disentangled variational auto-encoder (DVAE) alone, and the same DVAE associated with another classifier (Crecon). The DVAE alone exhibited an accuracy of 97.9%, and the DVAE + Crecon showed an accuracy of 99.8%. Sensibility and precision were also high for both sexes. These results are better than those reported from previous studies. These data-driven algorithms are easy to implement, since the pre-processing step is also entirely automatic. Fully automated methods save time, as it only takes a few minutes to pre-process the images and predict sex, and does not require strong experience in forensic anthropology.

Sex estimation of the adult Neandertal Regourdou 1 (Montignac, France): Implications for sexing human fossil remains

Sex is a biological trait fundamental to the study of hominin fossils. Among the many questions that can be addressed are those related to taxonomy, biological variability, sexual dimorphism, paleoobstetrics, funerary selection, and paleodemography. While new methodologies such as paleogenomics or paleoproteomics can be used to determine sex, they have not been systematically applied to Pleistocene human remains due to their destructive nature. Therefore, we estimated sex from the coxal bone of the newly discovered pelvic remains of the Regourdou 1 Neandertal (Southwest France, MIS 5) based on morphological and metric data employing two methods that have been recently revised and shown to be reliable in multiple studies. Both methods calculate posterior probabilities of the estimate. The right coxal bone of Regourdou 1 was partially reconstructed providing additional traits for sex estimation. These methods were cross validated on 14 sufficiently preserved coxal bones of specimens from the Neandertal lineage. Our results show that the Regourdou 1 individual, whose postcranial skeleton is not robust, is a male, and that previous sex attributions of comparative Neandertal specimens are largely in agreement with those obtained here. Our results encourage additional morphological research of fossil hominins in order to develop a set of methods that are applicable, reliable, and reproducible.

Measuring pelvises in 3D surface scans and in MDCT generated virtual environment: Considerations for applications in the forensic context

Virtual Anthropology (VA) transposes the traditional methods of physical anthropology to virtual environments using imaging techniques and exploits imaging technologies to devise new methodological protocols. In this research, we investigate whether the measurements used in the Diagnose Sexuelle Probabiliste (DSP) and Ischio-Pubic Index (IPI) differ significantly when 3D models of a bone are generated using 3D surface scans (3DSS) and Multidetector Computed Tomography (MDCT) scans. Thirty pelvises were selected from the SIMON identified skeletal collection. An equal ratio of females to males was sought, as well as a good preservation of the bones. The pelvises were scanned using an MDCT scanner and a 3D surface scanner. The measurements of the DSP and IPI methods on the dry bones (referred to as macroscopic measurements here), and then to the 3D models. The intra- and interobserver, using the Technical Error of Measurement (TEM) and relative Technical Error of Measurement (rTEM) error was assessed, and we aimed to observe if the measurements made on the MDCT and 3DSS generated models were significantly different from those taken on the dry bones. Additionally, the normality of the data was tested (Shapiro-Wilk test) and the differences in measurements was evaluated using parametric (Student t-tests) and non-parametric (Wilcoxon) tests. The TEM and rTEM calculations show high intra and interobserver consistency in general. However, some measurements present insufficient inter- and intraobserver agreement. Student t and Wilcoxon tests indicate potentially significant differences of some measurements between the different environments. The results show that especially in the virtual environment, it is not easy to find the right angle for some of the DSP measurements, However, when comparing the measurement differences between dry and virtual bones, the results show that most of the differences are less than or equal to 2.5 mm. Considering the IPI, the landmarks are already difficult to determine on the dry bone, but they are even more difficult to locate in the virtual environment. Nevertheless, this study shows that quantitative methods may be better suited for application in the virtual environment, but further research using different methods is needed.

Comparative study of sex estimates in adult skulls using direct measurement and tomographic image reconstruction

Sex estimation is an important procedure in forensic anthropology for human identification. The development of new technologies, such as three-dimensional computed tomography (CT), has provided excellent alternatives for this purpose. This study examined and compared a morphological method for sex estimation using two different approaches - direct measurement of physical structures and tomographic analysis using 3D images. A total of 111 skulls from the Museum of Human Anatomy at the University of São Paulo (MAH-USP) were used, (60 males and 51 females). All specimens were scanned by Philips Brilliance 64 CT scanner equipment and their corresponding images were reconstructed in three-dimensional (3D) models. The morphological characteristics of the skulls were analyzed by an observer who was blinded to the sex of the specimens. Five cranial structures were analyzed: external occipital crest, mastoid process, supraorbital margin, glabella, and mental eminence. The structures were scored 1 to 5 according to Buikstra and Ubelaker and validated by Walker. The success rates of the sex estimates obtained through direct measurement of the dry skulls ranged from 67.4% to 70.4% as compared to 60.2% to 68.1% for CT reconstruction. When analyzed separately, the maximum accuracy of the method was 68.33% in males and 88.24% in females in the physical analysis of structures. The glabella and mastoid process were the most effective structures to estimate sex through both techniques, respectively. Our results show that 3D CT images can be accurately used in the morphological analysis for sex estimation, representing a viable alternative in forensic anthropology.

Applying posterior probability informed thresholds to traditional cranial trait sex estimation methods

Sex estimation methods using traditional cranial nonmetric traits utilize predictive models to produce a final sex estimation, using the resulting model's score to classify the individual. When sex estimations are assigned from discriminant scoring alone, statistical confidence in the resultant estimate is not always assessed or reported. Although some forensic anthropologists may qualitatively report their confidence in the assessment (e.g., "probable male"), these statements are subjective, not standardized, and not necessarily based on statistical results in a uniform way. The goals of this study were to evaluate how posterior probability-informed thresholds (PPITs) impacted accuracy rates, assess the balance between sample inclusion and accuracy for the proposed PPIT approach, and make recommendations for the use and interpretation of specific thresholds in casework. Using a sample of U.S. Black and White females and males (n = 292), we examined how PPITs can standardize the decision-making process of inferring sex for two methods using nonmetric cranial traits. We found that using PPITs of at least 0.85 increased accuracy (over 92% for some PPITs) yet remained highly inclusive of the sample. PPITs < 0.75 did not produce classification accuracy rates significantly higher than chance, and when using these cranial trait sex estimation methods, cases with posterior probabilities (PPs) <0.75 should be reported as "indeterminate." The 0.75-0.84 PPIT interval had an accuracy rate of 76%, which was both statistically significantly different from chance as well as from the higher (>0.85) groups, suggesting that although sex estimation at this level may be acceptable, the results hold lower confidence.

A potential method for sex estimation of human skeletons using deep learning and three-dimensional surface scanning

Deep learning based on radiological methods has attracted considerable attention in forensic anthropology because of its superior classification capacities over human experts. However, radiological instruments are limited in their nature of high cost and immobility. Here, we integrated a deep learning algorithm and three-dimensional (3D) surface scanning technique into a portable system for pelvic sex estimation. Briefly, the images of the ventral pubis (VP), dorsal pubis (DP), and greater sciatic notch (GSN) were cropped from virtual pelvic samples reconstructed from CT scans of 1000 individuals; 80% of them were used to train and internally evaluate convolutional neural networks (CNNs) that were then evaluated externally with the remaining samples. An additional 105 real pelvises were documented virtually with a handheld 3D surface scanner, and the corresponding snapshots of the VP, DP, and GSN were predicted by the trained CNN models. The CNN models achieved excellent performance in the external testing using CT-based images, with accuracies of 98.0%, 98.5%, and 94.0% for VP, DP, and GSN, respectively. When the CT-based models were applied to 3D scanning images, they obtained satisfactory accuracies above 95% on the VP and DP images compared to the GSN with 73.3%. In a single-blind trial, a multiple design that combined the three CNN models yielded a superior accuracy of 97.1% with 3D surface scanning images over two anthropologists. Our study demonstrates the great potential of deep learning and 3D surface scanning for rapid and accurate sex estimation of skeletal remains.

Shape morphing technique can accurately predict pelvic bone landmarks

Diffeomorphic shape registration allows for the seamless geometric alignment of shapes. In this study, we demonstrated the use of a registration algorithm to automatically seed anthropological landmarks on the CT images of the pelvis. We found a high correlation between manually and automatically seeded landmarks. The registration algorithm makes it possible to achieve a high degree of automation with the potential to reduce operator errors in the seeding of anthropological landmarks. The results of this study represent a promising step forward in effectively defining the anthropological measures of the human skeleton.

Adjusting posterior probabilities to meet predefined accuracy criteria: A proposal for a novel approach to osteometric sex estimation

The osteometric methods are the most reliable way to estimate the sex of skeletons when DNA analysis is not used. However, as osteometric studies usually ignore the overlap in male and female skeletal dimensions, they rarely achieve accuracy sufficient for forensic application. To resolve this issue, recent studies suggest sex estimation only when posterior probability (pp) is greater than 0.95, but that approach does not always provide sufficient accuracy and creates a large proportion of unsexed skeleton. Thus, our study aimed to explore whether it is possible to adjust pp on skeletal measurements with pronounced sexual dimorphism to meet 95% accuracy and to enable sex estimation on a reasonable proportion of individuals. From 207 skeletons, we included 65 postcranial measurements and selected 10% of variables with the highest sexual dimorphism. We computed univariate and bivariate discriminant functions using pp threshold of 0.5, 0.95, and the threshold required to achieve accuracy of ≥ 95%. Discriminant functions with pp=0.5 obtained accuracy of 85%-93%, while those with pp≥0.95 and adjusted posterior probabilities obtained 94%-99%. However, we showed that by selecting a particular threshold, sex could be estimated on a greater proportion of individuals than for pp≥0.95: 42%-86% vs. 24%-62% for univariate and 69%-95% vs. 49%-78% for bivariate functions. Therefore, when developing sex estimation models, we suggest not to use fixed pp level, but to adjust pp to achieve 95% accuracy and to minimize the percentage of unsexed skeletons.

Impact of 3D surface scanning protocols on the Os coxae digital data: Implications for sex and age-at-death assessment

The 3D imaging technologies have become of paramount importance for example in disciplines such as forensic anthropology and bioarchaeology, where they are being used more and more frequently. There are several new possibilities that they offer; for instance, the easier and faster sharing of data among institutions, the possibility of permanent documentation, or new opportunities of data analysis. An important requirement, however, is whether the data obtained from different scanning devices are comparable and whether the possible varying outputs could affect further analyses, such as the estimation of the biological profile. Therefore, we aimed to investigate two important questions: (1) whether 3D models acquired by two different scanning technologies (structured light and laser) are comparable and (2) whether the scanning equipment has an effect on the anthropological analyses, such as age-at-death estimation and sex assessment. 3D models of ossa coxa (n = 29) were acquired by laser (NextEngine) and structured light (HP 3D Structured Light Scanner PRO 2) scanners. The resulting 3D models from both scanners were subjected to age-at-death analyses (via the quantitative method of Stoyanova et al., 2017) and sex analyses (via Diagnose Sexuelle Probabiliste 2 of Brůžek et al., 2017). Furthermore, high quality scans of a small sample (n = 5) of pubic symphyseal surfaces with the RedLux Profiler device were acquired as reference surfaces to which the outputs from both scanners were compared. Small deviations between surfaces were more evident in more rugged surfaces (in areas of depression and protrusion). Even though small differences from the reference surfaces were found, they did not have a significant effect on the age and sex estimates. It never resulted in the opposite sex assignment, and no significant differences were observed between age estimates (with the exception of those with the TPS/BE model).

CADOES: An interactive machine-learning approach for sex estimation with the pelvis

The pelvis is consistently regarded as the most sexually dimorphic region of the human skeleton, and methods for sex estimation with the pelvic bones are usually very accurate. In this investigation, population-specific osteometric models for the assessment of sex with the pelvis were designed using a dataset provided by J.A. Serra (1938) that included 256 individuals (131 females and 125 males) from the Coimbra Identified Skeletal Collection and 38 metric variables. The models for sex estimation were operationalized through an online application and decision support system, CADOES. Different classification algorithms generated high accuracy models, ranging from 85% to 92%, with only three variables; and from 85.33% to 97.33%, with all 38 variables. CADOES conveys a probabilistic prediction of skeletal sex, as well as a suite of attributes with educational applicability in the fields of human skeletal anatomy and statistics. This study upholds the value of the pelvis for the estimation of skeletal sex and provides models for that can be applied with high accuracy and low bias.

Sex estimation of os coxae using DSP2 software: A validation study of a Greek sample

Sex estimation methods based on skeletal remains vary on the selection of skeletal element, data acquisition and statistical approach resulting in variable classification accuracies that are highly dependent on the sample population. The only exception of this rule seems to be the os coxa, that appears to differ consistently between males and females across the globe. Currently sex estimation based on the os coxa can be easily estimated by taking ten measurements, inputting these in the DPS2 software and getting a sex estimate and the probability of correct group assessment. The performance of the software is highly reliable as confirmed by a validation study by Brůžek et al. (2017). Yet, there are still many populations not represented in the reference sample used to develop the software. The current study aimed to validate DPS2 using a sample from Crete, Greece. A total of 133 os coxae were measured following instructions on DSP2. Data were used to estimate sex with the software and to create population specific formulae for Greeks. DSP2 classified 117/133 (85.7%) of the sample with over 95% posterior probability (PP) of correct classification. Of the individuals classified with over 95% PP, only 3 were misclassified (2.6%).The best population specific formula only improved this percentage by 2.1% which indicates that DSP2 is a reliable tool for sex estimation in the Greek sample and it is recommended as method of choice in sex estimation of remains of unknown ancestry. If Greek ancestry is confirmed, population-specific formulae can be used in conjunction with DSP2 for a more reliable sex estimation.

Sex estimation using the mandibular canine index components

The mandibular canine index (MCI) has been described as a suitable methodology for sex estimation in forensic scenarios but there are contradictory reports about its accuracy. Moreover, the two mandibular canine teeth must be available, which is not always a viable option. The aim of this study was to strip the MCI by analyzing the MCI itself and its components, in order to optimize its use for sex estimation. The mesiodistal dimensions of the mandibular canine crown and the mandibular canine arch width were measured in a sample of 120 cast models. Five predictor variables were considered in this study: the standard MCI, a variation of the MCI using the left canine, and MCI components (MD43, MD33 and D33-43). Multivariate binary logistic regression was performed using stepwise forward approach to select the most statistical relevant variables on the probability of a cast being from a female. The estimated probability was then analyzed with respect to performance in sex classification (ROC analysis and optimal cut-offs accuracy) and compared with the performance of the univariate variables. MCI43 and MCI33 presented the lowest performance (64.2% and 63.3% respectively), and the highest overall accuracy was attained using the MD43 and MD33 (85.8% in both cases). The multivariate logistic model obtained (using MD43 and MD33) exhibited the same accuracy as the logistic model based solely on MD43 (85.8%). Our results suggest that MD43 should be used instead of MCI for sex estimation.

Application and validation of Diagnose Sexuelle Probabiliste V2 tool in a miscegenated population

The hip bone (os coxae) is the skeletal element that presents the greatest level of sexual dimorphism. Therefore, methods involving the analysis of the os coxae provide the most accurate sex estimation, and DSP2 (Diagnose Sexuelle Probabiliste v.2) is one of the most accurate tools used in this identification. The goal of this study is to apply and validate DSP2 in the identification of 103 os coxae (53 male and 50 female) belonging to a Brazilian-identified skeletal collection. Differences between sexes were statistically significant for all measurements, except for the acetabulo-symphyseal and spino-auricular lengths. From the 103 os coxae analyzed, there was a 9.43% error in male individuals and a 14% error in females. The results revealed that DSP2 can be applied to Brazilian-mixed populations with a good index of accuracy, although at a lower accuracy than other population samples. This study also clearly demonstrates that metric variation of the os coxae is extremely useful in sex estimation and reinforces the notion that pelvic sexual dimorphism is not population-specific.

Sex estimation standards for medieval and contemporary Croats

Aim

To develop discriminant functions for sex estimation on medieval Croatian population and test their application on contemporary Croatian population.

Methods

From a total of 519 skeletons, we chose 84 adult excellently preserved skeletons free of antemortem and postmortem changes and took all standard measurements. Sex was estimated/determined using standard anthropological procedures and ancient DNA (amelogenin analysis) where pelvis was insufficiently preserved or where sex morphological indicators were not consistent. We explored which measurements showed sexual dimorphism and used them for developing univariate and multivariate discriminant functions for sex estimation. We included only those functions that reached accuracy rate ≥80%. We tested the applicability of developed functions on modern Croatian sample (n = 37).

Results

From 69 standard skeletal measurements used in this study, 56 of them showed statistically significant sexual dimorphism (74.7%). We developed five univariate discriminant functions with classification rate 80.6%-85.2% and seven multivariate discriminant functions with an accuracy rate of 81.8%-93.0%. When tested on the modern population functions showed classification rates 74.1%-100%, and ten of them reached aimed accuracy rate. Females showed higher classification rates in the medieval populations, whereas males were better classified in the modern populations.

Conclusion

Developed discriminant functions are sufficiently accurate for reliable sex estimation in both medieval Croatian population and modern Croatian samples and may be used in forensic settings. The methodological issues that emerged regarding the importance of considering external factors in development and application of discriminant functions for sex estimation should be further explored.

Sex determination using the DSP (probabilistic sex diagnosis) method on the coxal bone: Efficiency of method according to number of available variables

The DSP method (probabilistic sex diagnosis) was applied to100 contemporary coxal bones from elderly individuals of the South of France. Ten variables with a posterior probability greater or equal to a 0.95 threshold were used. There was no statistical difference between right side and left side measurements. There was no mistake for sex assignment but the level of indetermination varied a great deal. It was higher in females than in males. The best combinations were obtained when using all 10 variables, some combination of 9 variables (all except SS or SIS or VEAC) or the first 8 variables. We conclude that the DSP method is of great interest in forensic anthropology, thanks to a very weak possibility of mistake when using the software for sex determination of the coxal bone. The main drawback is the level of indetermination that can be high depending on the available variables.

Metric Methods for the Biological Profile in Forensic Anthropology: Sex, Ancestry, and Stature

The biological profile, conducted by a forensic anthropologist, is necessary for severely decomposed or skeletonized remains. The biological profile consists of estimates of sex, age, ancestry, and stature. It is crucial to have a correct estimate of sex, as this designation will narrow down the search through missing persons reports by half (e.g., searching through NamUs). However, sex estimates can be population specific, necessitating accurate ancestry estimation. When estimates of age and stature are added, the search narrows further. If these estimates are incorrect, the unidentified human remains may never be identified. These biological profile components are estimated based on either metric or nonmetric methods (visual observation and recording of categorical data). While age is inherently nonmetric, stature is inherently metric. Estimates of sex and ancestry can take a metric or nonmetric approach. The purpose of this review article is to review metric methods in forensic anthropology (sex, ancestry, and stature), to provide general knowledge of why and how these metric methods work, and to highlight that estimates of sex, ancestry, and stature do not subscribe to a "one size fits all" model.

Geometric morphometrics reveals restrictions on the shape of the female os coxae

The methodology for sex determination in human skeletal remains depends on the different bone morphologies presented by men and women. Due to their direct implications in reproduction, the whole pelvis, particularly the os coxae, shows different characteristics in either sex. The sacrum and the os coxae constitute the birth canal. In this research study, the os coxae shape is analyzed using geometric morphometrics, providing information on morphology, regardless of size or any other factor beyond the geometry itself. A total of 46 adult ossa coxae from a Spanish archaeological collection were studied using geometric morphometrics. The results show that there is a restriction on the shape of female os coxae. In contrast, male os coxae presents a greater range of variation. The biological reason for this difference is the obstetrical dilemma; a concept defined as the anatomical conflict between bipedalism and the full-term birth of a neonate whose large head requires greater dimensions in the pelvic cavity. Our experimental data reinforce the validity of the obstetrical dilemma as source of the restriction on the shape of female ossa coxae. Additionally, according to the results obtained, size itself does not represent a condition for belonging to one sex or another.

A review of sex estimation techniques during examination of skeletal remains in forensic anthropology casework

Sex estimation is considered as one of the essential parameters in forensic anthropology casework, and requires foremost consideration in the examination of skeletal remains. Forensic anthropologists frequently employ morphologic and metric methods for sex estimation of human remains. These methods are still very imperative in identification process in spite of the advent and accomplishment of molecular techniques. A constant boost in the use of imaging techniques in forensic anthropology research has facilitated to derive as well as revise the available population data. These methods however, are less reliable owing to high variance and indistinct landmark details. The present review discusses the reliability and reproducibility of various analytical approaches; morphological, metric, molecular and radiographic methods in sex estimation of skeletal remains. Numerous studies have shown a higher reliability and reproducibility of measurements taken directly on the bones and hence, such direct methods of sex estimation are considered to be more reliable than the other methods. Geometric morphometric (GM) method and Diagnose Sexuelle Probabiliste (DSP) method are emerging as valid methods and widely used techniques in forensic anthropology in terms of accuracy and reliability. Besides, the newer 3D methods are shown to exhibit specific sexual dimorphism patterns not readily revealed by traditional methods. Development of newer and better methodologies for sex estimation as well as re-evaluation of the existing ones will continue in the endeavour of forensic researchers for more accurate results.