Abstract
BACKGROUND: X-ray grating interferometry has emerged as a promising imaging technique for breast computed tomography (BCT), offering complementary contrast from refraction beyond conventional absorption imaging. Although it provides higher resolution, the required dose levels remain above those typically used for breast cancer diagnostic imaging. PURPOSE: This study aims to design an optimized grating interferometry-based BCT system using a novel optimization metric that simultaneously maximizes signal strength from refraction and attenuation while considering grating fabrication and system constraints. METHODS: A systematic grid search was conducted to identify system configurations that optimize contrast-to-noise ratio while accounting for dose efficiency in both attenuation and refraction. The optimized system was benchmarked against a commercial absorption-based BCT system and a previously published grating interferometry design. Simulations were performed using in silico breast phantoms of varying diameters, incorporating realistic imaging conditions and noise modeling across dose levels from 1 to 100 mGy. Additionally, a signal fusion technique was applied to combine the two contrast channels. RESULTS: Quantitative analysis demonstrates that refraction provides superior contrast for small malignancies, while absorption remains beneficial for larger lesions. Fusing both contrast channels enhances the contrast-to-noise ratio and improves resolution in in silico reconstructed images, outperforming conventional absorption-based BCT at dose levels relevant for breast cancer diagnostic imaging. CONCLUSIONS: The findings suggest that an optimized grating interferometry-based BCT system can improve image quality by effectively combining absorption and refraction signals. This work underscores the need to balance imaging performance with practical implementation constraints, offering a framework that may extend to other imaging applications beyond breast cancer detection.