Digital versus screen-film mammography: a retrospective comparison in a population-based screening program

Digital radiography has several advantages over screen-film radiography in data storage and retrieval, making it an attractive alternative to screen-film radiography in screening mammography programs, if it performs as well.

We retrospectively compared screen-film mammography, photon-counting direct radiography, and computed radiography with population-based screening data from the Breast Unit at Helsingborg Hospital, Sweden, collected between January 2000 and February 2005. Outcomes were cancer detection rates, recall rates, and positive predictive values for breast cancer detection in women reappearing for screening.

Data were available for 52,172 two-view mammography examinations of 24,875 women. No initial screening (prevalence) examinations were included. Cancer detection rates based on mammographic findings were 0.31% (81/25,901) for film, 0.49% (48/9841) for photon-counting, and 0.38% (63/16,430) for computed radiography. The recall rate for film was 1.4%, which was significantly higher than that for PC-DR (1.0%; P<0.001) and computed radiography (1.0%; P<0.001). The positive predictive value was lower for film (22%) than for photon-counting (47%; P<0.001) and computed radiography (39%; P<0.001). In addition, the average glandular dose was 1.1mGy for film, 0.28mGy for photon-counting and 0.92mGy for computed radiography. Thus, photon-counting provided a 75% dose reduction, and computed radiography a 16% dose reduction, over film.

Digital radiography, especially photon-counting, performs as well as or better than screen-film radiography. Given the advantages related to improved data storage and communication, digital radiography seems to be a valid alternative to screen-film radiography.