Repeatability of a morphoscopic sex estimation technique for the mental eminence on micro-focus X-ray computed tomography models

Objectives: Walker's sex estimation method is based on five morphoscopic features of the human skull. This study aimed at testing the repeatability of one of the five traits, the mental eminence, visually on three-dimensional (3D) models, compared to the traditional, tactile scoring approach on bone. Materials and methods: The sample comprised 105 mandibles from the Pretoria Bone Collection and their respective virtual 3D models, obtained from micro-focus X-ray computed tomography (micro-XCT) scans. Four observers independently scored the bones first, followed by the virtual 3D modality. Intra- and interobserver errors (interOE and intraOE) were performed with Fleiss’ and Cohen's Kappa, respectively. We calculated the intermodality agreement per observer with Wilcoxon Signed-Rank tests. Results: The intraOE was moderate on bone (κ=0.448) and substantial on 3D (κ=0.799), while the Fleiss’ Kappa test for the interOE resulted in slight agreement both on bone (κ=0.163) and 3D (κ=0.169) irrespective of level of experience. All Wilcoxon Signed-Rank test P-values were significant. Discussion and conclusion: The application of the morphoscopic sex estimation for the mental eminence to micro-XCTs could be a matter of personal affinity as the level of experience did not play a role in the results. The expression of the mental eminence trait being population-specific, the individual's population affinity should be considered when sex is estimated in South African skeletons. It remains unclear whether the slight agreement between observers is due to the unreliability of the trait for sex estimation or whether the modalities are not easily interchangeable.