Population-Based Analysis of Female Thorax Dimensions and Body Habitus Classification Using CT Imaging
Keywords:
Keywords: Mammographic imaging, breast positioning, anatomical variability, thoracic dimensions, body habitus, CT imaging, Bayesian Network, Multiple Correspondence Analysis, rib measurements, body types, imaging quality, personalized mammography, early cancer detection.Abstract
Introduction: Mammographic imaging quality heavily depends on optimal breast positioning. Anatomical variability in female thorax size and body habitus can impact positioning accuracy. This study aims to establish reference ranges for thoracic dimensions in females using CT imaging and to categorize body habitus to guide mammographic positioning practices.
Methods: A retrospective analysis was conducted on 347 de-identified chest CT scans from females aged 40–89 years retrieved from the Medical Imaging and Data Resource Center (MIDRC) [17]. Thoracic measurements were performed digitally at six anterior rib landmarks. A Bayesian Network (BN) model was constructed to identify statistical relationships between rib measurements and body habitus categories. Multiple Correspondence Analysis (MCA) was used to classify participants into three body types [34,35].
Results: The BN model categorized body habitus into lean (20.5%), norm (55.6%), and curvaceous (23.9%) groups. Rib cage widths ranged from 115 mm to 126 mm across measured landmarks. Scenario analysis using BN allowed predictive estimation of habitus category from given rib measurements.
Conclusion: This study successfully quantifies thoracic size variations in a large female cohort and demonstrates how BN modeling can support personalized mammographic positioning. Integration of these findings into image evaluation criteria can improve imaging quality and early cancer detection.
