Investigating sexual dimorphism of the mastoid, bimastoid, occipital, and facial triangles in different craniofacial skeletal patterns using three dimensional models generated from CT scans.

We aimed to introduce a new approach to sex determination using 3D models of skull triangles including mastoid (MT), bimastoid (BMT), occipital (OCT) and facial (FT) triangles to explore how to achieve better accuracy, and also to assess the influence of skeletal malocclusion parameters (ANB, SNA), cephalic index (CI), and Welcher basal angle (WBA). The study was conducted on152 3D computed tomography scans, in which the dimensions of skull triangles, ANB, SNA, WBA, cranial length (CL) and breadth (CB) were measured. All parameters except the angular measurements were higher in males than in females. Univariate discriminant function analysis (DFA) yielded an average accuracies ranging from 54.6% to 79.6%. Multivariate DFA increased the percentage of correct predictions of gender to vary from 79.6 to 89.5%. Receiver operator characteristic curve (ROC) analysis showed that the area under ROC curve value to vary from 0.458 to 0.870. There was a general tendency for a decrease in all triangle dimensions except FT with increasing CI. In some cases, skeletal malocclusion, CI, and WBA influenced the area of triangles. In general, dolichocephalic individuals with larger WBA had lower RMT and LMT areas, whereas dolichocephalic and mesocephalic individuals with Class III had higher BMT and OCT areas. In FT, brachycephalic subjects with Class III had the lowest and Class I had the highest area value. Our results suggest that the MT, BMT, FT, and CL can be used effectively for sex determination and skeletal malocclusion, CI, and WBA had some influence in dimensions of skull triangles.