Application of Purkait´s triangle for sex prediction in digital postmortem CT scans

Abstract Estimating the sex of decomposed bodies and skeletal remains of unknown individuals is one of the first steps in the identification process of a forensic anthropological examination. As postmortem computed tomography (CT) scans are becoming a more common tool in forensic casework, the classical methods would be applied on these images. In this article, we applied the method called ´Purkait's triangle´ for sex estimation on CT scans. We hypothesize that the accuracy of sex estimation will be comparable to the application on dry bones. For this study, we acquired a total of 99 cases of postmortem CT scans from routine cases at the Institute of Forensic Medicine, Zurich, Switzerland. The femora were extracted using segmentation techniques. After this Purkait's method was applied on the segmented bone. Using the equations establish by Purkait we reached an overall accuracy of 67%. Using the method on the total length measurements showed an overall accuracy of 72% additionally a cutoff point could be established using a logistic regression. Despite some limitations, the results obtained by ´Purkait's triangle´ on CT scans for sex estimation in femoral bones showed high accuracies although not as high as applied in the classical way. Further research and testing different segmentation software might improve the result. Nevertheless, we were able to calculate a new cutoff point for a modern swiss population.