Morphoscopic observations in clinical pelvic MDCT scans: Assessing the accuracy of the Phenice traits for sex estimation in a Western Australian population

Abstract

The formation of a biological profile (including the estimation of sex and age) is usually the first task undertaken when skeletal remains are analysed by a forensic anthropologist. Recent literature attests to a growing awareness of the value of research focusing on age and sex related morphoscopic features visualized in high resolution multiple detector computerized tomography (MDCT) scans. The present study provides insight into the analysis of MDCT scans, with a specific focus on quantifying the accuracy of the Phenice sex estimation method in a Western Australian population. The sample comprises 448 clinical pelvic MDCT scans representing 226 male and 222 female individuals between 18 and 64 years of age. The scans (all ≤ 1.5 mm slice thickness) are reconstructed using three-dimensional volume rendered models in OsiriX software. A precision test was performed prior to data collection to quantify observer accordance; thereafter the accuracy of the identification of the Phenice pelvic traits (ischiopubic ramus; ventral arc; subpubic concavity) are statistically quantified. Intra-observer accordance is above 0.81 (Kappa value) for each morphological attribute assessed. Congruent with previous research the single most accurate trait overall is the ventral arc (86.61%), albeit an age-related sex-bias in classification accuracy was demonstrated. Based on the combined assessment of the three features, it is demonstrated that the Phenice method facilitates a high degree of expected accuracy in the classification of sex (92.24%), thus indicating that the method can be successfully applied in MDCT scans and is suitable for forensic application in a Western Australian population.