Abstract
PURPOSE: The breast shape differs between the prone position in breast magnetic resonance imaging (MRI) and the supine position on an operating table. We sought to determine the relationship between patient position-induced changes on prone and supine MRI in breast shape and mammographic breast density or thickness. MATERIALS AND METHODS: We evaluated data from 68 women with 69 breast cancers in this retrospective observational study. The difference in the minimal distance from the nipple to the pectoralis major (DNPp-s) or the internal thoracic artery between the prone and supine MRI (DNIs-p) was defined as the breast shape changes. Mammographic breast density was assessed by conventional 4-level classification and automated and manual quantification using a dedicated mammography viewer. The compressed breast thickness was recorded during mammography (MMG). We determined the association between patient position-induced breast shape changes on MRI and mammographic breast density or compressed breast thickness on MMG. RESULTS: On the conventional 4-level qualification, one breast appeared fatty, 39 appeared with scattered density, 23 appeared heterogeneously dense, and 6 breasts appeared extremely dense. Both automated and manual quantification of mammographic breast density differed between the 4 levels (p < 0.01 for both) and correlated with the 4 levels (p < 0.001 for both, r = 0.654 and 0.693, respectively). The manual quantification inversely correlated with DNPp-s and DNIs-p (p < 0.01 and < 0.05, r = - 0.330 and - 0.273, respectively). The compressed breast thickness significantly correlated with DNPp-s and DNIs-p (p < 0.01 for both, r = 0.648 and 0.467, respectively). CONCLUSION: Compressed breast thickness during MMG can predict the degree of patient position-induced changes in breast shape on MRI. The manual quantification of the mammographic breast density, which may reflect the biomechanical properties of the breast tissues, also correlates to the breast shape changes.