Author: Jeffrey S. Nelson, Raj Kumar Panta, Megan K. Russ, Ehsan Samei 👨🔬
Affiliation: Clinical Imaging Physics Group, Department of Radiology, Duke University Health System 🌍
Purpose: Contrast-enhanced mammography (CEM) enhances tumor detection by utilizing energy-dependent information from iodinated contrast agents. However, there is a lack of quantitative techniques to assess the impact of imaging variables such as tissue density, breast thickness, and imaging dose on iodine detectability. This study introduces the limit of detection (LOD) metric in CEM to quantitatively assess iodine detectability as a function of these imaging variables in a phantom study.
Methods: A CEM phantom simulating breast tissue was used, featuring a 10mm contrast target slab made of a 50/50 glandular-adipose mix with inserts representing iodine surface concentrations of 0.2, 0.5, 1, and 2mg/cm². CEM imaging was performed using a clinical protocol across four phantom tissue densities - BR-GLAND, BR-12, BR-30/70, and BR-FAT, each with tissue thicknesses of 20mm, 40mm, 60mm, and 80mm, added on top of the 10mm target slab. Additionally, imaging dose dependency was assessed at a 40mm thickness for each tissue density by varying the AEC settings. The LOD metric was estimated from CEM recombination images to assess iodine detectability.
Results: A strong linear relationship (R² > 0.99) was observed between iodine mean pixel values and the exponential of surface iodine concentration. The highest iodine detectability (lowest LOD) was observed at 60mm, with LOD of 0.77, 0.92, 0.87, and 1.00mg/cm² for BR-GLAND, BR-12, BR-30/70, and BR-FAT, respectively. Conversely, the lowest detectability (highest LOD) occurred at 80mm, consistent across all tissue densities. BR-GLAND exhibited superior LOD performance at lower thicknesses (20–60mm) but showed significant degradation at 80mm compared to lower-density tissues. Increasing the imaging dose improved LOD performance across all tissue densities.
Conclusion: This study demonstrates that LOD is strongly dependent on tissue density, thickness, and radiation dose in CEM imaging. The LOD metric can be used for optimizing CEM imaging protocols to enhance iodine detectability.