Sex Estimation From the Paranasal Sinus Volumes Using Semiautomatic Segmentation, Discriminant Analyses, and Machine...

Area of Science:

• Forensic Anthropology
• Radiology
• Anatomy

Background:

• Paranasal sinuses are air-filled cavities within the skull.
• Volumetric analysis of paranasal sinuses can reveal differences between sexes.
• Accurate sex estimation is crucial in forensic and anthropological studies.

Purpose of the Study:

• To investigate sex, age, and side-based differences in paranasal sinus volumes.
• To evaluate sexual dimorphism rates using discriminant function analysis and machine learning.
• To assess the accuracy of machine learning algorithms in sex determination based on sinus volumes.

Main Methods:

• Computed tomography (CT) images of 100 individuals were analyzed.
• Paranasal sinuses were segmented, and their volumes and densities were measured.
• Discriminant function analysis and machine learning algorithms were employed for sex determination.

Main Results:

• Males exhibited significantly larger mean paranasal sinus volumes than females (P < 0.05).
• No significant differences were found based on age groups or right-left sides (P > 0.05).
• Frontal sinus volume provided the highest accuracy for sex estimation; sphenoid sinus was least accurate.

Conclusions:

• Moderate sexual dimorphism exists in paranasal sinus volumes.
• Machine learning significantly improves sex estimation accuracy compared to traditional methods.
• Future studies combining linear, volumetric measurements, and machine learning may further enhance sex estimation rates.