Objectives: To estimate the degree of overdiagnosis of breast cancer in a mammographic screening programme.
Setting: A mammography service screening programme in Florence, Italy.
Methods: We studied the incidence of breast cancer in Florence between 1990 and 1999, following the introduction of screening in 1990. Incidence of breast cancer in this period was compared with incidence between 1985 and 1989, before the introduction of screening. It was necessary to estimate the number of cancers that would have arisen in the absence of screening, but after the end of followup (31 December 1999), so that these were not misclassified as overdiagnosed tumours. Around 60,000 women aged 50-69 were invited for screening during the period of study.
Results: There were 2780 breast cancers diagnosed during the period of study (2626 were invasive). There was no significant evidence of overdiagnosis of invasive cancers. When invasive and in situ cancers were considered together, around 5% of cases were overdiagnosed.
Conclusions: There is a small amount of overdiagnosis of ductal carcinoma in situ in mammography screening; however, this should not deter women from being screened. Training and practice in mammographic screening should emphasise detection of small, invasive lesions. Research into the natural history and treatment of the disease should aim at minimising overtreatment of those in situ lesions that are less likely to progress to invasive disease.