Assessing Sternal Dimensions for Sex Classification: Insights from a Greek Computed Tomography-Based Study.

Area of Science:

• Forensic Anthropology
• Medical Imaging
• Human Osteology

Background:

• Assessing the utility of sternal morphometric parameters from MDCT for sex estimation.
• Investigating their application in forensic identification within a contemporary Greek population.

Purpose of the Study:

• To evaluate sternal morphometric parameters derived from MDCT for sex estimation.
• To develop and validate statistical and machine learning models for forensic identification using these parameters.

Main Methods:

• Acquired sternal measurements from chest MDCT scans of 100 Greek adults (50 males, 50 females).
• Analyzed morphometric variables including sternum length, surface area, angle, index, manubrium dimensions, sternal body dimensions, and xiphoid process dimensions.
• Applied logistic regression and Random Forest classification to determine predictive accuracy.

Main Results:

• Both logistic regression (91% accuracy) and Random Forest (91% accuracy) models demonstrated high classification performance.
• Logistic regression identified manubrium body length (MBL) and sternal body length (SBL) as most predictive.
• Random Forest highlighted sternal surface area (SSA) as the most influential feature.

Conclusions:

• MDCT-derived sternal morphometry is a dependable, non-invasive method for sex estimation.
• Specific parameters like MBL, SBL, and SSA possess significant discriminatory power.
• Findings support the creation of population-specific standards for forensic use.