LAUSR

Purpose To compare image quality quantitatively and qualitatively of two dimension synthesized mammography (SM) images versus standard full- field digital mammography (FFDM) images in dedicated phantoms. Methods Two identical Hologic Selenia Dimensions mammography systems were considered, both with a dedicated software for the reconstruction of 2D synthetic images (C-View) from tomosynthesis projections. Analysis was made in terms of spatial resolution, contrast resolution, homogeneity and contrast-detail curve using specific phantoms: • TOR MAX (Leeds Test Objects, UK), associated with the automatic analysis system AutoPIA 3.6.6, • CDMAM 3.4 (Artinis, NL), associated with the analysis system CDMAM Analysis 1.5.5 (Euref), • AGATHA phantom (Leeds Test Objects, UK ), • CIRS model 011A (CIRS, USA), • a homemade phantom made of 25 aluminium spheres with a 1 mm diameter inserted in a 5 mm thick PMMA slab in a 24 × 24 cm area and 1 mm thick steel edge. The data were analysed with common software packages dedicated to quality control: IQworks, ImageJ, QC-DR and a homemade software. To make the comparison we always used processed images in MG format for both modality. Wherever possible, we used the clinical mode of acquisition (AUTO-Filter). Results Spatial resolution, both in terms of MTF and cut off resolution, was very different for the two modalities, with a clear superiority of the FFDM images; as to the contrast resolution, however, for the two modalities the difference found is minimal (Fig. 1). The contrast- detail curve again showed the clear superiority of FFDM images (Fig. 2). On SM images the presence of trailing artefacts is evident especially in the tube travel direction, showing a great effect (Fig. 3) Download : Download high-res image (502KB) Download : Download full-size image Download : Download high-res image (489KB) Download : Download full-size image Download : Download high-res image (582KB) Download : Download full-size image Conclusions This study show a substantial difference in the physical characteristics between SM and FFDM images, the latter showing a much higher resolution. SM images, however, can better highlight smaller objects, like microcalcifications, and actually increase their visualization.