Development of populational female thorax sizes and body habitus categories using computed tomography (CT) images

Abstract

INTRODUCTION: Positioning of the breast during mammography examination is critical to producing optimum quality images. However, the variation in female thorax and body habitus may affect mammography positioning for best image quality. This study aims to establish populational female thoracic (rib cage) sizes and quantify female body habitus categories. METHODS: A retrospective analysis of 347 female computed tomography (CT) chest axial scans was retrieved from an open access database to establish female rib cage sizes. Dimensions of the rib cage were measured digitally across six cross sections with six anterior rib landmark points and recorded in millimetres (mm). A Bayesian Network (BN) model was developed to establish the relationships of information extracted from the rib cage image data to quantify and categorise female body habitus. RESULTS: Female body habitus (thoracic size) were classified into three cohort categories: lean (20.5 %),norm (55.6 %) and curvaceous (23.9 %). The mean values (115–126 mm) and the corresponding 80 %prediction interval ranges for the healthy female rib cage size were obtained through the BN scenario analysis. CONCLUSION: The significance of this study is that it categorised 55.6 % of female thoracic sizes as being normal on a larger group of female population. This study contributed to a good understanding of the range of female body habitus (thoracic size) to improve positioning practice and maximize image quality. IMPLICATIONS FOR PRACTICE: This study used a BN model to establish a range of female thorax sizes using CT chest images to support improved positioning practices. These ranges of thorax sizes should be integrated as a body habitus criterion in the current image evaluation system to maximise image quality and subsequent breast cancer diagnosis.