Factors associated with false-positive screening mammography in São Paulo, Brazil.

There is limited data on the influence of various factors on mammography accuracy in low- and middle-income regions. In this cross-sectional study using registry data, we examined the sensitivity of screening mammograms by comparing mammography results with biopsies-confirmed breast cancer diagnoses in the state of São Paulo, Brazil. Additionally, we evaluate factors related to the screened women and to the breast lesions that could affect false-positive mammograms results. All screening mammograms conducted from January to December 2012 and biopsy results from January 2012 to December 2013 in the São Paulo State were retrieved from the Brazilian Breast Cancer Information System. We gathered details on women-related factors such as age, hormone therapy usage, prior radiotherapy, skin color, education level, skin type, breast density, and familial history of cancer, as well as on breast lesions, including type, size, characteristics, edges, and topographic site on the breast. To assess the risk effect of these factors on false-positive mammography results, we employed logistic regression analyses. Our results indicate that age under 50 years, use of hormone therapy, dense breasts, lesions smaller than 10 mm with defined edges, and the presence of calcifications were predictors of false-positive mammograms results. Finally, we observed that false-positives lead to longer times to diagnosis. These findings are relevant for the planning and management of organized breast cancer screening programs.