Effect of Transition to Digital Mammography on Clinical Outcomes

Purpose: To determine the effect of transition to digital screening mammography on clinical outcome measures, including recall rate, cancer detection rate, and positive predictive value (PPV). Materials and Methods: Institutional review board approval and the need for informed consent were waived for this HIPAA-complaint study. Practice audit data were obtained for three breast imaging radiologists from 2004 to 2009. These data were sorted by time period into the following groups: baseline (2004-2005), digital year 1 (2007), digital year 2 (2008), and digital year 3 (2009). The x 2 and Fisher exact tests were used to assess differences in proportions among and between years. Clinical outcomes based on lesion type from 2004 to 2008 were also compared. Computer-aided detection was used. Results: The three radiologists interpreted 32 600 screen-film mammograms and 33 879 digital mammograms. Recall rates increased from 6.0% at baseline to 7.1% in digital year 1 (P <.0001) and continued to increase in subsequent years to 8.5%. The cancer detection rate increased from 3.3 at baseline to 5.3 in digital year 1 (P =.0061), and it remained higher than that at baseline in subsequent years. PPV after screening mammogaphy (PPV(1)) increased from 5.6% at baseline to 7.5% in digital year 1 and returned to baseline levels in digital year 3. In contrast, PPV after biopsy (PPV 3) decreased from 44.5% at baseline to 30.3% in digital year 3 (P =.0021). From 2004 to 2008, 3444 patients with 3493 lesions were recalled. The percentage of recalls for calcifications increased from 13.8% at baseline to a peak of 23.9% in digital year 1 and 17.9% in digital year 2. Both PPV(1) and PPV(3) decreased for calcifications after the digital transition. Conclusion: Recall rate and cancer detection rate increase for at least 2 years after the transition to digital screening mammography. PPV 3 is significantly reduced after digital transition, primarily in patients with microcalcifications. (C) RSNA, 2011