Abstract
For quality control of breast cancer screening, it is essential to maintain an adequate recall rate and positive predictive value. In Yokohama, we apply a unique secondary film reading system for mammography screening. All screening mammograms are checked by several physicians. After the first diagnosis, screening mammograms are gathered for a secondary film reading service. If the first diagnosis is consistent with the second one, the diagnosis is adopted as the final diagnosis. However, if the first diagnosis differs from the second one, the final diagnosis is made by conference of 4~5 physicians. We examined whether our system is effective for achieving an adequate recall rate and positive predictive value.Subjects and methods: Our secondary film reading service dealt with 21,553 cases in 2006, and 25,650 cases in 2007. Method A: The final diagnosis is defined as positive irrespective of whether the first or second diagnosis is positive. Method B (our present method): A positive diagnosis is made when the secondary film reading service committee arrives at a positive diagnosis. We compared the recall rates and positive predictive values for methods A and B.Results: The recall rates and positive predictive values for method A were 8.8% and 2.1% in 2006, and 8.8% and 2.6% in 2007, respectively. The corresponding values for method B were 5.5% and 3.2% in 2006, and 5.8% and 4.0% in 2007, respectively. Method B showed a significantly lower recall rate and a higher positive predictive value than method A.Conclusion: Our secondary reading method is effective for maintaining adequate quality of breast cancer screening.
