U.S.-Norway Comparison of Interval Breast Cancers

Background: Publications from the International Breast Screening Network (IBSN) have shown that varying definitions create hurdles for comparison of screening performance. Interval breast cancer rates are particularly affected. Objective: to test whether variations in definition of interval cancer rates (ICR) affect comparisons of international ICR, specific to a comparison of ICR in Norway and North Carolina (NC). Methods: An interval cancer (IC) was defined as a cancer diagnosed following a negative screening mammogram in a defined follow-up period. ICR was calculated for women ages 50-69, at subsequent screening in Norway and NC, during the time period 1996 - 2002. ICR was defined using three different denominators (negative screens, negative final assessments and all screens) and three different numerators (DCIS, invasive cancer and all cancers). ICR was then calculated with two methods: 1) number of ICs divided by the number of screens, and ICs divided by the number of women-years at risk for IC. Results: There were no differences in ICR depending on the definition used. In the 1-12 month follow up period ICR (based on number of screens) were: 0.53, 0.54, and 0.54 for Norway; and 1.20, 1.25 and 1.17 for NC, for negative screens, negative final assessment and all screens, respectively: The same trend was seen for 13-24 and 1-24 months follow-up. Using women-years for the analysis did not change the trend. ICR was higher in NC compared to Norway under all definitions and in all follow-up time periods, regardless of calculation method. Conclusion: The ICR within or between Norway and NC did not differ by definition used. ICR were higher in NC than Norway. There are many potential explanations for the difference