[Show abstract] [Hide abstract]
Conversion of an ongoing screening program to digital mammography is studied in a pilot project. The purpose of this paper is to compare screening outcomes of digital and conventional screenings in the first year of the project.
METHOD AND MATERIALS
In a breast cancer screening center that carries out 70,000 screenings each year digital mammography was introcuced to study practical implications and quality issues related to digitization. A digital environment was set up that included a Full Field Digital Mammography (FFDM) system (Lorad Selenia), a dedicated mammographic review station (MeVis), a PACS with an archive and a technician workstation for viewing of current and prior mammograms (Rogan-Delft), and a CAD system with the capability of archiving digitized priors (R2 Technology). Women who were invited for a digital exam were randomly selected based on availability of the systems in use. In the first phase of the project only initial screenings were performed with FFDM. In the second phase also subsequent screenings were carried out digitally. Prior mammograms of one previous screening round were routinely digitized and archived for all participants on the day before they attended screening. These were viewed using automatic hanging protocols by the radiologists and by the technicians, as in the conventional screening protocol. Radiologists had CAD available. All cases were double read and in each reading at least one of a team of two experienced readers was involved. Five other radiologists took part in the double reading.
In the first year of the project 3959 women were screened with FFDM. Detection with FFDM was similar to detection with film. The referral rate for first screenings was significantly higher with FFDM than with film in the first phase of the project (0.58% vs 0.26%, p<0.01), but decreased to a normal level (0.27%) in the second phase. Also for subsequent screenings we found a significant increase in referral rate in the initial period (2.4% vs 1.2%). Comparison with prior mammograms was more difficult due to the way FFDM was processed.
Introduction of digital mammography had no significant effect on detection but initially resulted in a higher referral rate.
N.K.: Is shareholder and has received grant support from R2 Technology
Radiological Society of North America 2005 Scientific Assembly and Annual Meeting; 11/2005