Population-Based Assessment of Female Thoracic Dimensions and Body Habitus via CT Imaging

Abstract

Introduction: The quality of mammographic imaging is strongly influenced by precise breast positioning. Variations in female thoracic anatomy and body habitus can affect positioning accuracy, potentially impacting diagnostic outcomes. This study aimed to establish reference ranges for female thoracic dimensions using computed tomography (CT) and to classify body habitus patterns that can inform optimized mammographic positioning. Methods: A retrospective analysis was performed on 347 de-identified chest CT scans of females aged 40–89 years, obtained from the Medical Imaging and Data Resource Centre (MIDRC). Thoracic measurements were digitally recorded at six anterior rib landmarks. A Bayesian Network (BN) model was developed to explore statistical relationships between rib measurements and body habitus classifications. Multiple Correspondence Analysis (MCA) was applied to categorize participants into three body habitus types. Results: The BN model classified participants into lean (20.5%), norm (55.6%), and curvaceous (23.9%) habitus groups. Rib cage widths across the six measured landmarks ranged from 115 mm to 126 mm. Scenario testing with the BN model enabled predictive estimation of habitus category based on rib measurements. Conclusion: This study provides population-based reference data on female thoracic dimensions and demonstrates the utility of BN modelling in predicting body habitus. Incorporating these findings into mammographic positioning protocols may enhance image quality and facilitate earlier breast cancer detection.