Incidence of interval breast cancers after 650,000 negative mammographies in 13 Italian health districts

Objectives To estimate the total proportional incidence of interval breast cancers in a two-yearly mammography screening programme, and to perform subgroup analyses by woman's age, screening centre-specific recall rate and screening round. Methods Using unconverted electronic data-sets from the 13 screening centres in the Emilia-Romagna Region of northern Italy (540,450 women aged 50-69 years), a database of 919,538 mammography records was created. Of these, 655,175 eligible single-mammography records (1997-2002) from 379,318 women were record-linked with the regional Breast Cancer Registry. In the two-year inter-screening interval, a total of 1,022,694.3 woman-years at risk were accumulated, with 695 interval cancers observed and 2428.3 expected. The observed number of interval cancers was divided by the expected number to obtain the proportional incidence. Results The total proportional incidence of first- and second-year interval cancers was 0.18 (95% Cl 0.15-0.20) and 0.43 (0.39-0.47), respectively. Woman's age was inversely associated with proportional incidence in both interval years, with a cut-off point at age 60. A screening centre-specific recall rate greater than the regional average of 5% was associated with a proportional incidence of 0.14 (0.11-0.17) versus 0.20 (0.17-0.24) in the first interval year, and of 0.36 (0.31-0.41) versus 0.50 (0.44-0.56) in the second. The proportional incidence remained unchanged between the first and subsequent screening rounds. Conclusions The results were in line with the previous Italian data and with the recommended European standards. The inverse effect of woman's age and of recall rate was expected.