We developed a prototype photon-counting mammography unit with a cadmium zinc telluride detector, which provides a new type of image with physical analysis parameters. Using the X-ray attenuation information obtained from this device, we examined the ability of this technique in discriminating substances and estimating their compositions. To estimate the substance compositions, we used resected breast tissues immediately after a surgical operation for invasive carcinoma of no special type, and used phantoms to reproduce mammary glands and adipose tissue. In our system, the spectrum penetrating the substance was measured with three energy bins in each pixel. The products of linear-attenuation coefficient and thicknesses for each bin were calculated. Using these three values, the scatterplots displaying all the values calculated from each pixel inside the region of interest (ROI) on the image were created. The scatterplot displaying only gravity values calculated for each ROI on the image was created for evaluating the separation of plot points to discriminate between different substance compositions. The gravity points placed on the malignant tumor tissue were plotted separately from those on the normal tissue. Furthermore, a fusion image was created by overlapping an X-ray image and values of this scatterplot points represented on a 10-step color scale. The fusion image was highlighting the differences in substance compositions using color tone, such as malignant tumor or mammary gland tissue, by adjusting the color scale level.