Clinical outcome assessment in mammography: an audit of 7,506 screening and diagnostic mammography examinations

PURPOSE: To perform an audit of our routine mammographic practice and to compare our results to performance benchmarks. MATERIALS AND METHODS: We analyzed the outcomes of 7,506 consecutive examinations performed in 1 year. Screening and diagnostic cases were evaluated separately and mammographic assessments were based on the Breast Imaging Reporting and Data System (BI-RADS) classification. RESULTS: In 6,858 (91%) screening and 648 (9%) diagnostic cases, outcomes varied substantially. The recall rate was 10.9%. Estimated sensitivity and specificity were similar (100% vs. 98% and 88% vs. 94%) in the screening and diagnostic groups. Positive predictive values (PPV1, PPV2, and PPV3) were higher in the diagnostic group compared to the screening group (64%, 65%, and 68% vs. 4.9%, 33%, and 39%, respectively). Cancer outcomes in the screening and diagnostic groups were, respectively, as follows: cancer detection rate, 6.1 per thousand vs. 86.4 per thousand; mean invasive cancer size, 15.7 mm vs. 24.5 mm; minimal cancers, 38% vs. 19%; stage 0-1 cancers, 50% vs. 21%; and lymph node negativity, 76% vs. 29%. CONCLUSION: The measures of our screening outcomes were concordant with the literature and the performance benchmarks for screening mammography; however, in our diagnostic group, the reasons for the higher PPV, higher cancer detection rate, and the diagnosis of cancer in a more advanced stage compared to the performance benchmarks should be investigated with more detailed periodic audits.