Anti-scatter grid performance in digital mammography and contrast-enhanced mammography: a Monte Carlo study

Abstract

Objective. Digital mammography (DM) and dual-energy contrast-enhanced mammography (CEM) use the same anti-scatter grid, despite their differences in x-ray spectra. This study investigated grid performance under clinically relevant conditions in DM and CEM across representative ranges of patient characteristics and assessed the effect of increased grid ratio r. Approach. Monte Carlo simulations of low-energy (LE) and high-energy (HE) images of a clinical DM/CEM system were performed using realistically shaped compressed breast phantoms with varying thickness, size, and composition. Linear grids with r= 5, 10, 15, and 20 were simulated. Grid performance was measured via the signal-difference-to-noise-ratio (SDNR) improvement factor (SIF), and differences in spatial scatter distributions were assessed via scatter-to-primary ratio (SPR) profiles. Main results. Breast composition had no considerable impact on grid performance for DM or CEM. SIF increased with breast size by an average of 1.3% (LE) and 6.9% (HE), and with breast thickness by 21%–30% (LE) and 19%–27% (HE). The standard grid (r = 5) did not reduce SDNR in LE or HE images for the examined thickness range (30–90 mm). Higher grid ratios improved SPR homogeneity between LE and HE images in the inner projected breast area. A grid with r = 10 yielded up to 4% higher SIF at the centre of mass than r = 5 in HE images, while causing a maximum SDNR loss of 3.5% for 30 mm thick breasts in LE images. Significance. This study provides clinically relevant measures of grid performance in DM and CEM, closing a gap on missing insights of breast size effects. Contrary to previous findings, SDNR is not necessarily degraded for thin breasts under standard imaging conditions. Increasing the grid ratio to r = 10 improves HE image quality and might reduce rim artifacts in CEM due to increased SPR homogeneity, with minimal SDNR loss in DM.