Sex estimation from morphometry of lumbar vertebrae (L4-L5): A machine learning approach

Area of Science:

• Forensic anthropology
• Bioinformatics
• Medical imaging

Background:

• Sex estimation is crucial for biological identity in forensic investigations.
• Pelvic and cranial bones are standard for sex estimation, but vertebrae are used when these are unavailable.
• This study investigates the utility of fourth and fifth lumbar vertebrae for sex identification in the Turkish population.

Purpose of the Study:

• To evaluate the effectiveness of fourth and fifth lumbar vertebrae in sex estimation using machine learning.
• To identify key morphometric parameters for sex determination in these vertebrae.
• To assess the applicability of these methods within the Turkish population.

Main Methods:

• Retrospective analysis of computed tomography (CT) images from 540 individuals (aged 18-79).
• Collection of eight specific measurements from the fourth and fifth lumbar vertebrae.
• Application of eight distinct machine learning algorithms for sex estimation.

Main Results:

• Logistic regression achieved the highest accuracy: 91.3% for the fourth lumbar vertebra and 90.7% for the fifth.
• Vertebral Body Anterior-Posterior Diameter and Vertebral Body Anterior Height were the most significant parameters for sex estimation.
• The study confirms sexual dimorphism in the fourth and fifth lumbar vertebrae.

Conclusions:

• Fourth and fifth lumbar vertebrae are viable for sex estimation, particularly in the Turkish population.
• Machine learning models, especially logistic regression, demonstrate high accuracy.
• Findings support the use of lumbar vertebrae morphometry in forensic anthropology and clinical practice.