Prediction of the soft-tissue facial morphology matrices in a European population, a Geometric Morphometric study using Cone- Beam Computed Tomography.

Abstract

The increasing number of unidentified and unclaimed bodies at state mortuaries in Gauteng, the smallest yet most densely populated province in South Africa, has become a cause for concern for the Health Department (Sobuwa, 2023). The socioeconomic backgrounds of these victims pose challenges in the identification process, as the usual methods like dental records or DNA/fingerprint comparisons are often unavailable. This is primarily due to the high cost of DNA testing and the lack of regular dental care among most South Africans, resulting in a lack of comparative records. Consequently, alternative approaches such as craniofacial reconstruction need to be explored to facilitate the identification process. The purpose of this research was to examine the impact of various factors, such as population affinity, sex, age, and allometry, on the variation in craniofacial shape. Additionally, the study aimed to investigate the correlation between the morphology of hard and soft tissues in the face using cone-beam computed tomography (CBCT) scans, and to develop prediction models based on these findings. For this study, a total of 76 white South Africans and 108 French nationals, with ages ranging from 18 to 80 years, were included. The CBCT scans utilized in this research were obtained retrospectively from two different sources. Specifically, the CBCT scans of the South Africans were collected from the University of Pretoria's Oral and Dental Hospital as well as the Groenkloof Life Hospital in South Africa. On the other hand, the CBCT scans of the French nationals were obtained from the University of Bordeaux in France. Landmarks were selected on both the facial skeleton (including orbits, nasal bones, anterior nasal aperture, zygoma, and maxilla) and the soft features of the face (ears, eyes, nose, and mouth) to create regions of interest. A total of 43 craniometric landmarks, including 559 sliding landmarks, were registered on the hard-tissue surfaces, while 50 capulometric landmarks were registered on the soft-tissue surfaces using the MeVisLab© v. 2.7.1 software. Geometric morphometric techniques were employed to analyze the craniofacial shape differences associated with population affinity, sex, age, and allometry. Furthermore, two- blocks partial least squares (PLS) analyses were used to assess the covariation between the matrices of hard and soft tissues in each sample. Finally, the Projection onto Latent Structures Regression (PLSR) algorithm was utilized to develop predictive statistical models based on the obtained data.

The analysis of the physical structure of both hard and soft tissues in the facial region can aid in identifying the degree of connection between the underlying skull and the soft facial features. This can lead to a more comprehensive understanding of the factors that contribute to the variations in craniofacial morphology among diverse populations. The interrelation between the bony structures of the face and the soft facial features is crucial for the development of craniofacial reconstruction techniques, and comprehending this association can facilitate the creation of precise predictive models. The shape variation of the facial skeleton and soft facial features were influenced by population affinity, sexual dimorphism, ageing, and allometry, according to the findings. The impact of population affinity was the most significant (p-value < 0.05), with population- specific differences observed between white South Africans and the French. The effects of sexual dimorphism and ageing differed between the two populations, with specific hard and soft tissue elements affected by these variables. Additionally, the correlations between hard and soft tissue matrices were notably different between the two populations. Overall, the study highlights the importance of considering population affinity, sex, and age when reconstructing the face and emphasizes the need for accurate prediction models specific to different populations.

The PLSR method was utilized in this study to create reliable predictive models, which were further improved by incorporating additional parameters such as population affinity, sex, and age. The study found that age and sex were crucial factors to include as supplementary information in order to enhance prediction performance. The analysis was conducted on 184 specimens, and the precision of the estimated soft-tissue shape of the ears, eyes, nose, and mouth was evaluated using metric deviations on both trained and untrained datasets. The trained data showed that the prediction errors for the ears, eyes, nose, and mouth for white South Africans were 2.299 mm (SD 0.132), 2.069 mm (SD 0.154), 2.523 mm (SD 0.145), and 3.573 mm (SD 0.003), respectively, while prediction errors for the French group were 2.693 mm (SD 0.064), 2.248 mm (SD 0.040), 2.643 mm (SD 0.165), and 3.585 mm (SD 0.003). On the other hand, prediction errors on non-trained data were slightly greater, with errors of 3.038 mm (SD 0.132), 1.976 mm (SD 0.154), 3.266 mm (SD 0.317), and 3.774 mm (SD 0.301) reported for white South Africans' ears, eyes, nose, and mouth, respectively, whereas prediction errors of 3.432 mm (SD 0.064), 2.155 mm (SD 0.040), 3.386 mm (SD 0.317), and 3.572 mm (SD 0.294) were recorded for the French group.

The present study aimed to assess the effects of population affinity, sex, age, and allometry on the shape components of facial matrices composed of hard and soft tissues. The findings of this investigation demonstrated notable variations in craniofacial morphology between white South Africans and French individuals. These results underscore the importance of comprehending and quantifying the impact of these factors, as it can significantly improve the accuracy of prediction models when applied to specific populations. Ultimately, this research contributes to a better understanding of the intricate relationship between population characteristics and facial shape.