Comparison of automatic and manuel readings for CDMAM phantom images

Abstract

In this study, we present comparison of human and software reading of contrast-detail mammography (CDMAM) phantom images for a digital mammography system. The human readings were conducted through CDMAM images using 3 Megapixel and 5 Megapixel medical monitors. The phantom was positioned between Poly methyl methacrylate (PMMA) blocks of 20mm thickness. A set of 8 digital radiographs were taken using the automated exposure control (AEC-Opdose). The phantom was slightly repositioned between each exposure, always with the smaller details close to the chest wall side. Automated readings were carried out with CDMAM 3.4 Analyser software V.2.8. (CDCOM) and human readings were evaluated by three observers. As a result of this study, CDMAM 3.4 Analyser software is able to provide results significantly better than those estimated by human observers. Additionally, human readings were better for 5 Megapixel medical monitor. We can note that quality control tests of mammography systems and clinical evaluation of mammogram should be evaluated with 5 Megapixel medical monitor.