| Abstrakt: |
The femur is one of the most commonly recovered bones in mass casualty disasters and is often used for identification, both for height and sex. However, in most cases the femur cannot be obtained as a whole. Therefore, this study evaluated and compared the reliability of sex estimates based on measurements taken from the proximal and distal ends of the femur. The study was conducted by evaluating the measurements of 225 femurs on radiological images obtained from a total of 128 individuals (67 males and 61 females). All the radiological images were obtained from the archive of the Radiology Department of Baskent University nine of the anthropometric measurements were linear and two were angular. When the anthropometric measurements obtained from the whole sample were evaluated in terms of differences between the right and left sides, significant differences were observed between the sides in terms of intercondylar angle, intertrochanteric distance (bc), inclination angle (inc), and femoral head width (b - b). On the other hand, as a result of statistical analyses performed to predict sex, significant differences were found between both sexes in all measurements except intercondylar angle and inclination angle. Logistic regression analysis was performed and a formula was developed to determine sex with the data obtained. A ROC analysis was performed to determine the discriminability threshold of the measurements taken from the distal and proximal ends of the femur for sex estimation. The findings show that transverse head diameter (E-E), femur length (C-C), femoral head width (A-A), and femoral neck width (B-B) measurements are reliable parameters with an accuracy rate of 97.0%, 94.6%, 93.5%, 91.6%, respectively, and the epicondylar width taken from the distal end is the most reliable parameter with an accuracy rate of 98.1%. A LOOCV (confusion matrix) was performed on all measurements and the validated model was found to be 91% successful in predicting sex according to logistic regression analysis. [ABSTRACT FROM AUTHOR] |
