Metric analysis of the postcranial skeleton: a comprehensive approach for biological sex estimation in an Italian population.

Abstract

Objectives: This paper presents a metric methodology for estimating biological sex specifically tailored to the Italian population. The method considers 121 standard metric measurements derived from 46 bones across various post-cranial regions.

Materials and methods: The sample consists of 400 individuals (M = 200; F = 200) from the 20th century CAL Milano Cemetery Skeletal Collection aged 20 to 104 years old. The sample was divided into a training subset (75%; n = 300) and a testing subset (25%, n = 100). Intra- and inter-observer analyses, as well as univariate sectioning points, and multivariable logistic regression analyses were performed.

Results: Intra- and inter-observer analysis showed excellent reproducibility of the measurements, with some exceptions generally related to the measurement of long bone diameters. Univariate sectioning points resulted in 18 measurements with accuracies exceeding 90%, and another 48 measurements achieving over 80% accuracy. In total, 43 multivariable logistic regression models were developed for 32 bones, and these models further increased the accuracy.

Discussion: The validation of these models demonstrated that the proposed methodology allows for sex estimation with accuracies of over or near 90% and minimal class discrimination bias across all post-cranial skeletal regions. The highest accuracies - with both sectioning points and multivariable models - were the radius (96.8%), scapula (95.3%), and tibia (95.2%). This study introduces a comprehensive metric standard for the Italian population and highlights the accuracy of the metric approach for estimating biological sex.