RESUMO
PURPOSE: For almost 3 years, Radiologists and Physicists from Padova and Ferrara Universities have collaborated together, with the aim of collecting and comparing experimental data useful to define the most significant parameters for quality controls in digital mammography. Successively, radiologists and physicists working in other sites where a digital mammography unit was installed joined the work-group. MATERIALS AND METHODS: In this study we report the results obtained from measurements of linearity, uniformity, short- and long-term reproducibility, AEC stability performed on 5 digital equipment by using a simple test object. X-ray beam quality and tube yield were preliminarily checked in such a way that possible uncertainties of digital system responses could be separated from those due to differences among X-ray tubes. RESULTS: Results showed that the equipment considered, comparable in terms of both beam quality (HVL) and tube yield, always displayed linear response and reproducibility errors lower than 5%. Uniformity was very good and the grey level compensation as a function of exposure parameters remained within 5%. Differences in choice of parameters by exposure control system (AOP) were emphasised, especially for crossing between track/filter combinations (from Mo/Mo to Mo/Rh and from Mo/Rh to Rh/Rh); those differences were attributed to the +/-2 mm tolerance of breast thickness measurement (mechanically obtained) greater than the AOP tolerance (+/-1 mm). DISCUSSION AND CONCLUSIONS: Obtained results can be useful, as comparison and reference values, for users employing a digital mammography unit of the same kind reported in this paper. Moreover, the same results could be used as "orientation" also by other users having different digital mammography technologies, whose operation should be nevertheless specifically studied and understood in order to find the most useful parameters for quality controls. The acquired experience clearly showed us that years of investigations will be necessary in order to be able to write reliable protocols. This should induce people to contemplate the necessity of not improvising "theoretical" protocols, that are unreliable and dangerous for their negative clinical implications.