Is there sufficient evidence that the hyoid bone can be used for sexual dimorphism? a systematic review and meta-analysis

Abstract

Objective: Sex estimation using the hyoid bone is a complementary approach in forensic anthropology, valuable when other skeletal elements are unavailable. Its morphology shows sexual dimorphism and age-related ossification and may assist neck-trauma assessment. However, population variability and methodological heterogeneity limit generalization. This systematic review and meta-analysis evaluate the accuracy and forensic applicability of hyoid-based sex estimation. Method: Following PRISMA guidelines, a systematic search was conducted in PubMed, Scopus, and Web of Science up to August 31, 2024. Human studies estimating sex from hyoid morphology were included, using macroscopic or imaging analyses with individuals of known sex as reference. Non-human studies, pathologies, and reviews were excluded. Reporting completeness and risk of bias were appraised with STROBE and QUADAS-2. A narrative synthesis and meta-analysis using random-effects models provided pooled sensitivity and specificity estimates. Results: Of 965 records, 23 studies were included, comprising 4,770 participants. Evidence supported sexual dimorphism of the hyoid, and fusion increased with age. Studies showed moderate-to-high reporting completeness but frequent bias concerns, particularly due to unclear patient-selection and reference-standard domains. The highest accuracies ranged from 69.2% to 93% using discriminant analysis and reached 95.4% with machine-learning approaches. A meta-analysis of six studies (827 participants) yielded pooled sensitivity 0.742 (95% CI 0.514–0.887) and specificity 0.799 (95% CI 0.612–0.909). Conclusion: The hyoid bone can be useful for sex estimation, especially in limited skeletal preservation. Multivariate and artificial intelligence models show promising performance, but evidence remains limited by heterogeneity. Future studies should standardize procedures and broaden validation across populations.