The identification of victims involved in mass fatality incidents, as well as the identification of unknown individuals in criminal cases has become an increasingly important issue nowadays. Sex assessment represents a key point in forensic evaluations due to its significance in providing biological identity. Even though the availability of documented skeletal remains to forensic practitioners is a common practice in many countries, in Turkey, contemporary documented skeletal remains are not available for this purpose. For this reason, studies have been focused on living populations. Previous research has shown that modern technologies such as CT scanning present very promising potential in establishing new standards for contemporary populations. Therefore, the main aim of this project was to examine the application of the measurements taken from 3D CT images of the femur in order to assess sex, and to contribute to the establishment of discriminant function equations for the Turkish population for forensic applications. The accuracy and reproducibility of imaging methods in the assessment of the measurements taken from femora are essential when estimating sex. This research also concentrated on determining the accuracy and repeatability of CT measurements, using the femur. Prior to primary data collection, a preliminary study was performed in an effort to test the reliability of the femur measurements. The results of reliability analysis indicated no significant difference between the three observations of each measurement. Thus, the methodology employed in the current study appears reliable and reproducible. In addition, a validation study was conducted to determine the linear measurement accuracy of the 3D volume rendering models derived from a medical CT scanner and the influence of different reconstruction parameters. The differences between measurements obtained from dry bones and their 3D volume rendered models were also evaluated. The results from this study indicated that there were no statistically significant differences between measurements taken from different reconstruction parameters and measurements obtained from CT images and drybones. Using the CT data, volume-rendering function (VR), 3D Curved Multiplanar reconstruction (MPR), and Scout View on OsiriX were employed in order to compare the accuracy and reliability of each rendering method and to determine which technique is optimal for linear measurements. Overall, the measurements taken from the 3D Volume Rendering images had the highest intra-observer reliability when compared to the other two rendering methods. This research study produced data and interpretations that will inform on and improve population specific standards of sex assessment from three-dimensional postcranial osteometric landmarks. Additionally, this research is believed to provide value for a developing discipline of forensic anthropology, and integrate within the existing systems of criminal investigation and disaster victim identification practices in Turkey. A Turkish sample population, consisting of 300 adult hospital patients was examined via the interpretation of CT reconstructed images using the OsiriX software. The 3D reconstructions were then created using the volume-rendering function in OsiriX (v.5.6.). Following the 3D reconstruction, an image of each femur was segmented from the surrounding bones to ensure the correct usage of landmarks as accurately as possible. Thirteen measurements were acquired using a 3D viewer after being located and marked on each CT reconstructed femora. These thirteen anthropometric parameters were measured and analysed by basic descriptive statistics and discriminant analysis methods using the SPSS 21.0 software package. The intra-observer variation was assessed by obtaining the intraclass correlation coefficient in order to evaluate the accuracy of the linear measurements taken. Asymmetry was also tested. The results indicated that an accuracy of 92.3% was acquired from a combination of six of the measurements, and the Femur Vertical Diameter of Neck (FVDN) measurement was found to be the most dimorphic with 88.0% accuracy.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:720974 |
| Date | January 2017 |
| Creators | Gulhan, Oznur |
| Contributors | Harrison, K. |
| Publisher | Cranfield University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://dspace.lib.cranfield.ac.uk/handle/1826/12272 |
Page generated in 0.0129 seconds