Estimates of overdiagnosis of breast cancer with mammography screening differ between studies. The estimate of overdiagnosis obtained using the cumulative incidence approach with data from randomized clinical trials ranges from 10 to 14%. In observational studies, where statistical adjustment for lead-time was utilized, the estimate was lower than 5% in two Italian studies, while it was quite higher in a Swedish study. Most of the differences are explained by different methodology, study design and measures. In this review we discuss the concept and the definition of the measure of overdiagnosis, and present different study methods to estimate overdiagnosis of breast cancer with mammography screening.